PB 222 001
COMBUSTION PRODUCTS FROM THE INCINERATION
OF PLASTICS
Michigan University
Prepared for
Environmental Protection Agency
] 97 3
DISTRIBUTED BY:
Ki
National Technical Information Service
II $ DEPARTMENT OF COMMERCE
V
-------
BIBLIOGRAPHIC DATA
SHEET
1. Report No.
EPA-670/2-73-049
2- | 3-^R.tf ipient's Accession No.
PB 222-001
4. Title and Subtitle
COMBUSTION PRODUCTS FROM THE INCINERATION OF
PLASTICS
5. Report Date
1973-issuing date
6.
7. Author(s)
E. A. Boettner, G. L. Ball, and B. Weiss
8- Performing Organization Rept.
No.
9. Performing Organization Name and Address
University of Michigan
School of Public Health
Environmental and Industrial Health
Anr. Arbor, Michigan 48104
10. Projcct/Task/Work Unit No.
11. Contract /Grant No.
EP-00 386
12. Sponsoring Organization Name and Address
U.S. Environmental Protection Agency
National Environmental Research Center
Office of Research & Development
Cincinnati, Ohio 45268
13 I'ype oi Report & Period
( overed
Final
14.
15. Supplementary Notes
Analys is of the combust ion products of plastics was undertaken for three reasons;
to provide scientists and engineers with information needed to design incinerators
in order to maximize their efficiency while minimizing maintainance and pollution;
to identify products of incomplete combustion potentially recoverable for their fuel
or crude chemical value; and to identify products of incomplete combustion which
would be acutely toxic in an accidental fire. Plastics studied were polyvinyl chlo-
ride, polysulfone, polyurethanes, polyimide, LopacR, BarexR, phenol formaldehyde,
urea formaldehyde, polyethylene, polypropylene, polystyrene, polycarbonate, poly-
pher.ylene oxide, polyester, synthetic fabrics (DacronR, Orion", nylon), and natural
products (wood and wool). One- to three-gram samples were heated at a controlled rate
from 5 to 50 C/min in the presence of a measured flow of air or air plus oxygen.
By this method plastics were never completely combusted to carbon dioxide and water,
but rather generated large numbers of gaseous and condensed products. Additional
gaseous products included straight-chain saturated and unsaturated hydrocarbons
through hexane, aromatic hydrocarbons, hydrogen chloride^ sulfur dioxide, cyanides,
ammcnia, and oxides of nitrogen. Liquefied fractions produced by most plastics
were complex mixtures of 10 to 50 compounds, including heterocyclic and polycyclic
hydrocarbons .
17. Key Words and Document Analysis. 17o. Descriptors
*Combustion, *Flastics, ^Incinerators, Air pollution, Maintenance,
Identifying, Fuels, Chemical properties, Toxicity, Fires, Oxygen,
Samples, Heat, Air, Condensing, Hydrocarbons, Polyvinyl chloride,
Polyethylene, Polypropylene, Polystyrene, Synthetic fibers, Natural
fibers
17b. Identif iers/Open-Ended -erms
Solid waste disposal, Resource recovery, Gaseous products, Polysulfone,
Polyurethanes, Polimide, Polycarbonate, Polyphenylene oxide, Poly-
ester, LopacR, BarexR, Phenolformaldehyde, Urea formaldehyde
17c. COSATl Field/Group 13-B, 11-E
18. Availability Statement
Release to public
19. Security Class (This
Report)
UNCLASSIFIED
21. No. of Pages
20. Security Class (This
Page
UNCLASSIFIED
22. Price
THIS FORM MAY BE REPRODUCED
-------
REVIEW NOTICE
The Solid Wa;:e Research Laboratory of the National
Environmental Research Center - Cincinnati, U.S. Environmental
*
Protection Agency, has reviewed this report and approved its
publication. Approval does not signify that the contents
I
necessarily reflect the views and policies of this laboratory
or of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute
endorsement or recommendation for use.
The text of this report is reproduced by the National
Environmental Research Center, Cincinnati, in the form re-
ceived from the Grantee; new preliminary pages have been
•supplied .
*
ii
-------
FOREWORD
Man and his environment must be protected from the
adverse effects of pesticides, radiation, noise and other
forms of pollution, and the unwise management of solid
waste. Efforts to protect the environment require a
focus that recognizes the interplay between the com-
ponents of our physical environment—air, water, and
land. The National Environmental Research Centers
provide this multidiscip1inary focus through programs
engaged in
• studies on the effects of environmental
contaminants on man and the biosphere, and
• a search for ways to prevent contamina-
tion and to recycle valuable resources.
In an attempt to investigate the problems of solid
waste disposal and pollution, this study analyzed com-
bustion products of plastics. Special consideration was
given to technical information and incomplete combustion.
A. W. Breidenbach, Ph.D.
Director
National Environmental
Research Center, Cincinnati
iii
-------
ABSTRACT
Analysis of tne combustion products of plastics was undertaken
for three reasons: to provide scientists and engineers with
information needed to design incinerators in order to maximize
their efficiency while minimizing maintainance and pollution, to ,
identify products of incomplete combustion potentially recoverable
for their fuel or crjde chemical value; and to identify products hf
incomplete combustion which would be acutely toxic in an accidental
fire. Plastics studied were polyvinyl chloride, polysulfone, /
polyurethanes, polyimide, Lopac^, Barex*, phenol formaldehyde, /
urea formaldehyde, polyethylene, polypropylene, polystyrene, poly-
carbonate, polyehenylene oxide, polyester, synthetic fabrics I
(DacronR, Orion**, nylon), and natural products (wood and wool)!.
One-to three-gram samples werfe heated at a controlled rate fron 5
to 50 C/min in the presence of a measured flow of air or air /1 us
oxygen. By this method plastics were never completely combusted
to carbon dioxide and water, but rather generated large numbers
of gaseous and condensed products. Additional gaseous products
included straight-chain saturated and unsaturated hydrocarbc/is
through hexane, aromatic hydrocarbons, hydrogen chloride, syI fur
dioxide, cyanides, ammonia, and oxides of nitrogen. Liquef/ed
fractions produced by most plastics were complex mixtures c/r 10
to 50 compounds, including heterocyclic and polycyclic hydibcarbons.
-------
TABLE OF CONTENTS
Abstract iv
Figures yii
Tables xii
Conclusions 1
Recommendations 2
Introduction 3
Methodology 6
Polyvinyl Chloride 19
Polysulfone 36
Polyurethanes 47
Polyimide 57
LopacjJ 63
BarexK 68
Urea Formaldehyde 73
Phenol Formaldehyde 76
Polyethylene 80
Polypropylene 87
Polystyrene 90
Polycarbonate 94
Polyphenylene Oxide 106
Polyester 123
Synthetic Fabrics (DacronR, Orlon^, Nylon) 124
Natural Products (Wood and Wool) 132
Acknowledgments 135
References 136
Publications Resulting From Research 140
v
-------
LIST OF FIGURES
number page
1 Schematic diagram of a cfifferential thermal 7
analysis (DTA)apparatus
2 Schematic diagram of a thermogravimetric 9
analysis (TGA) apparatus
3 Combustion furnace and temperature 10
control equipment
4 Mass spectrometer system (Adapted from 13
Roboz, J. Introduction to mass spectrometry.
New York, Interscience, 1968. p.10)
5 AEI MS10 mass spectrometer with glass 15
reservoir inlet
6 AEI MS30 mass spectrometer with gas 16
chromatograph, probe, and heated reservoir
inlets
7 Differential thermal analysis record of 21
polymer C heated at 10 C/min in air
8 Thermogravimetric analysis record of 22
polymer C heated at 3 C/min in air
9 Chromatogram of low-boiling combustion 23
products of polymer A on a Porapak Q
column
10 Chromatogram of combustion products of 24
polymer A on a column of 5 percent squalane
on Chromosorb P
11 Benzyl chloride production by two meat- 30
wrap, films (A and B) heated slowly to 200 C
12 Benzyl chloride production by two meat-wrap 31
films (A and B) heated slowly to 337 C
13 Benzyl chloride production by two meat-wrap 33
films (A and B) heated at 200 C
14 Benzyl chloride production by two meat-wrap 34
films (A and B) heated at 337 C
Preceding page blank
VI 1
-------
number page
15 Bisphenol A-polysulfone 37
16 Differential thermal analysis records of 38
polys-ilvone heated at 10 C/min in helium
(brr'-.en curve) and air (smooth curve)
17 Thermogravimetric analysis records of 39
polysulfor.e heated at 5 C/min in helium
(broken curye) and air (smooth curve)
18 A portion of an infrared spectrum of 40
polysulfone combustion products (10-cm
path-length gas eel 1)
19 Chromatogram of polysulfone liquid residue 41
on.a low K1 Durapak column
20 Differential thermal analysis record of 48
a polyurethane foam heated at 10 C/min in
ai r
21 Thermogravimetric analysis record of a 49
polyurethane foam heated at 10 C/min in air
22 Thermogravimetric analysis record of 50
polyurethane Sample E heated at 10 C/min
in air
23 Polyimide 58
24 Thermogravimetric analysis record of 59
polyimide heated at 10 C/min in air
25 Infrared spectrum of polyimide combustion 60
gas (10-cm path-length gas cell)
26 Differential thermal analysis records of 64
LopacR heated at 10 C/min in helium
(broken curve) and ai'r (smooth curve)
27 Thermogravimetric analysis record of 65
Lopac^ heated at 10 C/min in air
28 Infrared spectrum of Lopac^ combustion 66
gas (10-cm path-length gas cell)
29 Thermogravimetric analysis record of 69
BarexR heated at 10 C/min in air
vi i i
-------
30
31
32
33
34
35
36
37
38
39
40
41
42
" 43
iijuje
70
71
74
75
77
78
81
82
83
88
91
92
95
96
Infrared spectrum of Barex^ combustion
gas (10-cm path-length gas cell)
Chromatogram of Barax^ liquid residue on a
3 percent SE 30 column
Thermogravimetric analysis record of urea
formaldehyde heated at 10 C/min in air
Infrared spectrum of urea formaldehyde
combustion gas (10-cin path-length gas cell)
Thermogravimetric analysis record of
phenol formaldehyde heated at 10 C/min in
ai r
Infrared spectrum of phenol formaldehyde
combustion gas (10-cm path-length gas cell)
Thermogravimetric analysis record of low-
density polyethylene (powder) heated at
10 C/min in air
Thermogravimetric analysis record of high-
density polyethylene (powder) heated at
10 C/min in air
Thermogravimetric analysis record of high-
density polyethylene (pellets) heated at
10 C/min in air
Thermogravimetric analysis record of
isotactic polypropylene heated at 10 C/min
in air
Thermogravimetric analysis record of
polystyrene heated at 10 C/min in air
Chromatogram of polystyrene liquid
residue on a 3 percent SE 30 column
Bisphenol A-polycarbonate
Differential thermal analysis records of
polycarbonate heated at 10 C/min in helium
(broken curve) and air (smooth curve)
i x
-------
number page
44 Thermogravimetric analysis records of 97
polycarbonate heated at 5 C/min in helium
(broken curve) and air (.smooth curve)
45 Chrcmat^r^ii of polycarbonate liquid 99
vesiduL; on a low K' Durapak column
46 Poly-2,6-dimethyl -1 ,4-phenylene oxide 107
47 Differential thermal analysis records of 108
polyphenylene oxide heated at 5 C/min in
helium (broken curve) and air (smooth curve)
48 Thermogravimetric analysis record of 109
polyphenylene oxide heated at 5 C/min in
ai r
49 Differential thermal analysis records of 111
modified polyphenylene oxide heated at
5 C/min in helium (broken curve) and air
(smooth curve)
50 Thermogravimetric analysis record of 112
modified polyphenylene oxide heated at
5 C/min in air
51 Chromatogram of polyphenylene oxide 113
combustion products on a Porapak Q column
52 Chromatogram of polyphenylene oxide 114
combustion products on a column of 5 percent
Squalane on Chromosorb P
53 Chromatogram of polyphenylene oxide liquid 116
residue on a low K' Durapak column
54 Differential thermal analysis record of 125
DacronR heated at 5 C/min in air
55 Tiiermogravimetric analysis record of 126
Dacron^ heated at 10 C/min in air
56 Differential thermal analysis record of 127
0rlonR heated at 5 C/min in air
57 Thermogravimetric analysis record of 128
OrlonR heated at 10 C/min in air
x
-------
Differenti«il Lhcirmal analysis record of
nylon heated «it !3 C/min in air
Thermogravimetric analysis record of
nylon heated at 10 C/min in air
Chromatogram of wood combustion products
on a Porapak Q column
xi
-------
List of Tables
number page
1 Plastics Production, 1971 4
2 Identification of PoTyvinyl Chloride 25
Chrcratcgram Peaks
3 Comparison of Combustion Products of the 27
Plastics with the Combustion Products of
their Polymers
4 Variation of Combustion Products of 28
Polymer A with Temperatur
5 Identification of Polysulfone Residue 43
Chromatogram Peaks
6 Combustion Products of Polysulfone at 44
Several Combustion Conditions (800 C
maximum)
7 Combustion Products of Polysulfone at 45
Several Combustion Conditions (1000 C
maximum)
8 Polysulfone Combustion Products During 46
Several Temperature Ranges
9 Combustion Products of Polyurethane Foams 52
A and B at Several Combustion Conditions
10 Combustion Products of Polyurethane Foams 53
C and D at Several Combustion Conditions
11 Combustion Products of Polyurethane Sample 54
E at Several Combustion Conditions
12 Combustion Products of Polyurethane Sample 55
F at Several Combustion Conditions
13 Identification of Polyimide Residue 61
Combustion Products
14 Carbon Dioxide and Carbon Monoxide from 61
Polyimide Combustion
15 Cyanide and Nitrogen Oxides from 62
Polyimide Combustion
Xll
-------
16
17
18
19
20
21
22
23
24
25
26
27
28
29
me_
67
72
79
84
85
86
89
93
.100
101
102
104
117
118
Cyanide from Lopac'* Combustion
Tentative Identification of Barex^
Residue Chromatogram Peaks
Cyanide from Phenoi Formaldehyde
Combustion
Combustion Product^ of Low-Density
Polyethylene (Powder) at Several Com-
bustion Conditions
Combustion Products of High-Density
Polyethylene (Powder) at Several
Combustion Conditions
Combustion Products of High-Density
Polyethylene (Pellets) at Several
Combustion Conditions
Combustion Products of Isotactic
Polypropylene at Several Combustion
Condi tions
Identification of Polystyrene Residue
Chromatogram Peaks ,
Identification of Polycarbonate Residue
Chromatogram Peaks
Combustion Products of Polycarbonate at
Several Combustion Conditions (800 C
maximum)
Combustion Products of Polycarbonate at
Several Combustion Conditions (1000 C
maximum)
Polycarbonate Combustion Products During
Several Temperature Ranges
Identification of Polyphenylene Oxide
Residue Chromatogram Peaks
Quantities of Combustion Products from
Several Polyphenylene Oxide Plastics
x i i i
-------
number page
30 Variation of Polypheny!ene Oxide Combus- ^9
tion Products with Air Supply
31 Effect of Heating Rate on Combustion 121
Products of Polyphenylene Oxide
32 Variation of Polyphenylene Oxide Combus- 122
tion Products with Temperature
33 Identification of Wood Chromatogram Peaks 134
xiv
-------
CONCLUSIONS
The potential hazard from compounds generated on combustion of a
plastic depends on the primary structure of the polymer (its
atomic composition), the additives used in formulating the plastic,
and the conditions under which it is burned. Little is known of
the effect of additives, as the present study deals mainly with
combustion products of the polymer. Data accumulated to date in-
dicate that three categories of polymers should be considered:
those consisting of carbon, hydrogen, and oxygen; nitrogen-con-
taining polymers; and polymers with halogen or sulfur heteroatoms.
Plastics composed of only carbon and hydrogen or carbon, hydrogen,
and oxygen form carbon dioxide and water when completely combusted.
Incomplete combustion results in production of carbon monoxide as
the major toxicant, plus gaseous and condensed hydrocarbon products.
The condensate has significant fuel or crude chemical value but
may be a source of-polycyclic hydrocarbon pollution, particularly
from aromatic polymers. Plastics containing nitrogen as a hetero-
atom produce on complete combustion molecular nitrogen and small
amounts of oxides of nitrogen, as well as carbon dioxide and water.
On incomplete combustion hydrogen cyanide, cyanogen, nitriles
and ammonia may form in addition to hydrocarbon gases, presenting
a significant health hazard in open burning or an accidental fire.
Any liquid condensate formed may be composed of a variety of organic
nitrogen compounds as well as hydrocarbons. Nitrogen compounds
are more sensitive than other combustion products to changes in
combustion conditions. Generally the more incomplete the combustion,
the more ammonia and cyanide will form. Plastics containing halo-
gen or sulfur heteroatoms form acid gases such as hydrogen chloride,
hydrogen fluoride, and sulfur dioxide on complete combustion in
addition to carbon dioxide and water, and can form organic halogen
or sulfur compounds on incomplete combustion. These compounds
present.air pollution, incinerator corrosion, and toxicity problems
requiring special techniques to overcome.
1
-------
RECOMMENDATIONS
There are dozens of different commercial polymers and to each may
be added plasticizers, stabilizers, flame retardants, and other
chemicals to produce formulated plastics. It is impractical to
study the combustion products of each of the thousands of formula-
tions used, and so some generalizations have to be made. Such
generalizations can best be made from data on a few representative
samples of each class of polymers and each class of additives, and
then by studying known formulations to see if any synergistic
processes are occuring. It is recommended that studying combustion
products of representative polymers and additives individually
will give more meaningful and more widely usable results than
studying "synthetic rubbish" mixtures formulated in the laboratory
or rubbish mixtures with known amounts of plastics added. Either
of these methods tends to dilute and mask any specific effects of
specific products. Once analytical data are available for a poly-
mer (or plastic) it is possible to extrapolate that data by com-
puter modeling to a wide variety of applications, although syner-
gistic effects are certainly possible. Compiling present analyt-
ical data into a single source and filling in where data are
inadequate would provide a useful framework for predicting both
burning characteristics and toxicity of combustion products of
various plastics.
In view of the types of compounds identified in the liquid residues
after incomplete combustion of some plastics, we strongly recom-
mend a study be undertaken to determine if plastics can form poly-
cyclic hydrocarbons such as benzpyrene and benzanthracene on com-
bustion, and if formed, whether antipollution devices on inciner-
ators are sufficient to prevent them from entering the atmosphere.
Compounds of the same molecular weight as these carcinogens were
detected by mass spectrometry in the present study, but deficien-
cies in reference spectra prohibited positive identification of
the compounds found. It would not be surprising to find polycyclic
compounds generated in a pyrolysis or combustion process.
Further study on incineration of plastics should emphasize the
contribution of additives to combustion products, the composition
and economic value of liquid residues produced under pyrolysis
conditions, and generation of particulate matter and adsorption
of chemicals onto particulates. Study of gaseous products such as
hydrocarbons and cyanides, but not acid gases, are of lesser
importance in incineration since such compounds are generally
broken down on secondary burning. However, these gaseous products
may be acutely toxic and are of the utmost importance in open
burning or an accidental fire.
2
-------
INTRODUCTION
There are two basic environmental concerns involving plastics.
The first has to do with their disposal, and the second with the
toxicity of their products of combustion when burned (either in-
tentionally for disposal or accidentally as in home fires). We
began working on the latter problem about 1963 to determine what
hazardous products could be evolved in the heating of polymers
and plastics, to provide information to those concerned with the
health of people exposed to such products in closed or confined
spaces.
In 1969 the study was expanded to include all products of combus-
tion, toxic and non-toxic, in order to fill the need for this type
of information in the field of solid waste disposal. The amount
of solid waste disposed of by burning is continually increasing,
and as a result there is interest in designing incinerators to
improve their efficiency, to use the heat energy developed, and to
do as much as possible to eliminate them as secondary polluters
of air and water. It is thus necessary for the designer to know
more about the material he is trying to burn. This is especially
true of plastics. Plastics currently account for approximately 2
percent of the total packaging solid waste. They are projected to
account for 8 percent of the packaging solid waste by 1976 and 2.8
percent of the total solid waste by 1980.1,2
There is reason to believe the percentage of plastics in refuse
will increase as they continue to become more prevalent in furni-
ture, clothing, automobiles, construction materials, and packaging.
For example, Modern Plastics magazine recently speculated that a
two-billion pound per year market for acrylonitrile-based plastics
could result if plastic bottles now being test-marketed are
accepted for only two major brands of soft drinks.3 Plastics for
disposal are received in bulk by many municipal facilities from
manufacturers in their area.
Plastics production figures for 1971 (Table 1) have been reported
by Modern Plastics.^ Such figures were used in choosing repre-
sentative polymers for the study. The proposed goal of this
research grant was to examine combustion products of polysulfone,
polycarbonate, methyl methacrylate, polystyrene, acrylonitrile-
butadiene-styrene, polyethylene and polypropylene. In addition,
although information on polyvinyl chloride had already been pub-
lished, study of this plastic was to be extended to determine
products generated at higher temperatures. During the grant per-
iod it became apparent that there was greater interest in some
plastics not initially chosen for study. Polvurethanes were thus
substituted for methyl methacrylate and Lopac* and Barex* were
substituted for acrylonitrile-butadiene-styrene. This selection
3
-------
TABLE I
PLASTICS PRODUCTION, 1971*
Plastic
Million pounds
Polyethylene (high and low density)
6,400
Polystyrene (including copolymers)
3,840
Polyvinyl chloride (including copolymers)
3,320
Polypropylene
1,260
Phenolic
1 ,067
Urethane foam
930
Polyester
810
Urea and melamine
679
All non-PVC vinyls
650
A1 kyd
610
Coumarone-indene-and petroleum resins
360
Cellulosics
165
Miscellaneous
794
*Adapted from The Statistics for 1971, Modern Plastics,
49 (1): 41, January, 1972.
included some of the high-volume plastics as well as newer plastics
with good growth potential.
The combustion products of many polymers have been analyzed under
vacuum or inert atmosphere, primarily for the purpose of deter-
mining polymer structure. Madorsky describes such work done under
vacuum on fourteen types of polymers.^ No comprehensive work has
been done, however, on the combustion of these polymers or formu-
lations of them in air. Principle publications are listed in the
references. A group at the New York University College of
Engineering and Science, under the direction of E. R. Kaiser and
A. A. Carotti, studied the effect of plastics on the operation of
municipal incinerators in the New York City metropolitan area,
under the sponsorship of both the U. S. Department of Health,
Education, and Welfare and the Society of the Plastics Industry.6
Two Environmental Protection Agency-sponsored projects simultaneous
to this studied environmental effects of plastics combustion
products. The study, "Incineration of Plastics Found in Municipal
Wastes", directed by R. W. Heimburg of Syracuse University Corp.,
used a small model incinerator to generate products to which plants
and animals were exposed.' Some of the products were identified.
A study by Batelle, Columbus Laboratories, "Fireside Metal Wastage
in Municipal Incinerators", attempted to determine the mechanisms
involved in incinerator corrosion."
4
-------
This report includes the <|ua I i l.ili ve and quantitative information
on combustion product.', of (hose plastics studied by this labora-
tory. Some overlap, clearly indicated in the text in the poly-
vinyl chloride and polyphenylene oxide chapters, occurs with work
done under a previous grant. Since polyvinyl chloride is the
plastic most often of concern in solid waste management, its
inclusion seemed mandatory.
5
-------
METHODOLOGY
Introduction
The basic analytical system for identifying volatile combustion
products of plastics was developed before initiation of this grant
and has been modified over the years as new techniques were devel-
oped and new instrumentation added to the laboratory.^ The anal-
ysis of each polymer or plastic was carried out in three phases:
1. Thermal analysis was used to determine the temperatures
at which chemical and physical changes occurred in the
sample.
2. The sample was combusted in a small furnace and compounds
liberated within temperature ranges determined to be sig-
nificant by thermal analysis were identified or character-
i zed.
3. The quantities of identified products were determined,
either over a single decomposition step or over the
entire combustion run. •
Thermal Analysis
Differential thermal analysis (DTA). Differential thermal analysis
was used to determine the temperatures at which chemical and phy-
sical changes took place in a sample. The apparatus used (Figure
1) was a Robert L. Stone KAH unit. It consists of two chambers
which are heated at a continuous and identical rate; each chamber
contains a fast-response thermocouple, with the two couples con-
nected in series opposition. The sample (200 mg) is placed in one
chamber(x), and an inert reference material, such as aluminum
oxide, is placed in the other(s). As the chambers are heated,
generally at the rate of 5 to 10 C a minute, there is no output of
the series couples until either an endothermic or an exothermic
reaction takes place in the sample. In the first case, the
absorbed heat results in a slightly cooler sample chamber, which is
indicated by the differential couple. In the second case, the
sample chamber is slightly warmer than the reference, with a re-
sultant reversed polarity in the signal from the differential
couple. It will be noted that there are provisions for circulating
air at variable rates through each of the chambers. The chambers
have a volume of about 0.1 cc, although microchambers are available
that handle considerably less. A cylindrical furnace is placed
around the chamber housing, and the temperature of the furnace is
programmed to rise at a predetermined rate controlled by a tempera-
ture programmer. An amplifier is used for increasing the small
thermocouple signals to the millivolt level. The record obtained
using this apparatus shows the heat absorbed (from an endothermic
6
-------
Furnace
Cavities
Sample holder
Differential
couple
Standard
couple
Figure 1. Schematic diagram of a differential thermal analysis (DTA) apparatus
-------
change) or liberated (from an exothermic change) as aT, plotted
against the temperature of U=e sample. The aT-ful1-scale sensi-
tivity is from 0.5 to 3.5 C lor the various records presented in
the text.
Thermogravi metric analysis (TliA). rheriiiocjrdviiiietric analysis pro-
vided a record of weight change of the sample as it was heated.
Inasmuch as the DTA equipment contained the basic components (a
temperature controller and programmer) for a TGA unit, the remain-
der of the apparatus was built at the University of Michigan fol-
lowing a design developed at Dow Chemical Company by Cobler and
Miller.10 It consists (Figure 2) .of a small cylindrical furnace
containing a sample pan suspended on a quartz spring. In both
this apparatus and the DTA unit, the sample area is a gastight
chamber with complete control over the amount of entering air or
other gas and means for collecting all emerging gases. The sam-
ple is heated at a rate of 3 to 10 C per minute, and its tempera-
ture is recorded on the X-axis of an X-Y recorder. The displace-
ment of the platinum pan holding the sample, as measured by the
differential transformer below the quartz spring, is recorded on
the Y-axis. The final record is then a plot of percent weight
remaining versus sample temperature.
Sample Combustion
Combustion gases could be quantitatively collected from the TGA
apparatus, but it was found that larger quantities than were
available from this unit were desirable or essential. A combus-
tion furnace (Figure 3) with controlled temperature and air supply
was designed and built to carry out the combustion on as much as a
tenfold scale (2 g) but under the same conditions as in the TGA
apparatus. Samples (0.5-2 g) for qualitative and quantitative anal-
ysis were placed in porcelain boats in the VycorR gastight furnace
core. A measured amount of purified tank air was swept through
the core and collected, along with volatile products, in a SaranR
bag so that the total sample volume was usually 2 to 15 liters.
Air supplies were determined for each plastic by calculating the
amount necessary to completely convert all the carbon to carbon
dioxide when integrated over the entire combustion run, and using
an amount less than, roughly equal to, and more than that calcu-
lated. The amount calculated was, of course, a deficient supply
because all the air passing through the tube was not in the vic-
inity of the sample and thus available for reaction, and the reac-
tion would only occur with reasonable efficiency at the high end of
the temperature range.
The furnace was controlled by a West SCR Stepless Controller. The
sample was heated rapidly from room temperature to a temperature
8
-------
QUARTZ ROD
QUARTZ SPRING
<=n
DIFFERENTIAL
TRANSFORMER
jTAYL
QUARTZ ROD
WATER JACKET
FURNACE
PLATINUM CRUCIBLE
-THERMOCOUPLE
Figure 2. Schematic diagram of a thermogravimetr'ic analysis (TGA) apparatus
9
-------
-------
50 to 100 C below its decomposition temperature and then heated at
5 to 50 C per minute. Initially the upper temperature limit of
the apparatus was 800 C, but this temperature was not sufficient
to combust some of the newer high-temperature polymers in a reason-
able length of time. Thus the limit was changed to 1000 C.
Products formed were swept out of the heating zone rather quickly
by the continuous air flow, minimizing secondary reactions. Higher-
boiling combustion products were thus not broken down further and
condensed as liquid residue in either the cool end of the combus-
tion tube or the sample bag.
We believe our method of combusting the samples is realistic when
compared with the situation in a fire or incinerator for these
reasons: In most fires and in incinerators, the plastic is not
normally in an open flame, or if so, only toward the end of its
combustion process. For example, in a structural fire, the flame
will progress or spread toward the plastic (floor covering), so
that it is gradually heated before being exposed to the flame, and
then only if the flame proceeds that far. Likewise, in most incin-
erators the feed mechanism is such that the refuse is gradually
carried toward the flame (by conveyer or feed screw) at a rate that
makes for slow heating before reaching the flame to avoid cooling
the flame. In this case, of course, many of the combustion products
that we show will break down further into their combustion products
when they reach the flame, but one can frequently predict which of
these will do so.
Qualitative Analysis
Major combustion products, present in quantities ranging from 0.1
to 15 percent, were identified directly by infrared spectroscopy.
Minor combustion products were identified by mass spectrometry
after concentration and separation from the mixture.
Infrared spectroscopy. Combustion gas directly from the sample
bag was analyzed by infrared spectroscopy on a Perkin-Elmer 221,
in either a 10-cm path-length gas cell or a long-path (up to 40
meters) gas cell. Each compound has a characteristic pattern of
infrared absorption bands, allowing major components (>1000 ppm)
of the sample to be easily identified. This procedure of direct
infrared analysis was, however, not applicable to the minor com-
pounds of the combustion products for two reasons: (a) the large
number of lesser components would produce additive, overlapping
spectra, which would be impossible to resolve, and (b) the volume
of each component is insufficient for the practical application of
the long-path gas cell, normally used for the identification of a
component present in low concentration.
Gas chromatography. To separate minor components for the purpose
11
-------
of Identification, the gas chromatographic technique was used.
Gas chromatography involves a column packed with liquid-coated
solid support or active solid through which helium or other inert
gas is continuously flowing. When a sample is injected, it is
swept along by the helium; but components of the sample have dif-
ferent affinities for the liquid coating or active solid and so
are absorbed and desorbed at different rates. Thus, they take
different times to pass through the column, resulting in a sepa-
ration. The column 1s kept in an oven and the temperature can be
constant or programmed at any desired rate to elute compounds in
a reasonable length of time. A detecting device is located at the
end of the column to measure each component as it appears, the
output of the detector being recorded. The retention time of a
compound on the column is the only clue to its identification,
and exact reproducibility of retention times is difficult for a
sample producing a large number of closely-spaced peaks. Thus gas
chromatography 1s a good separation technique but a poor identi-
fication technique.
Several gas chromatographs were used throughout the study:
Research Specialties with argon ionization detector
Beckman GC-2A with hydrogen flame and thermal conductivity
detectors
Wilkens Model H600-B with hydrogen flame and electron
capture detectors
Microtek MT 220 with hydrogen flame detectors
Carle Basic Model 9000 with hydrogen flame detectors
Sampl e con centra ti on ¦ Samples were collected for identification
using the MT 220 chromatograph which has dual columns and detec-
tors. To prevent sample destruction by the detector, one column
has a stream splitter which allows 10 percent of the eluted gas to
go to the detector and 90 percent to emerge at a side arm for
collection. Concentration compensated for the reduced amount of
sample being detected and also reduced the number of times an
individual peak had to be collected before sufficient sample was
available for identification. A tenfold concentration was accom-
plished by passing the combustion gas through a U-tube immersed in
Methyl Cellosolve^-dry ice (-50 C) and then vaporizing the frozen
products by heating the U-tube and sweeping it with a small volume
of air, so that the total sample volume range was reduced from
2000 to 5000 cc to 200 to 500 cc.
Mass spectrometry. Many of the minor combustion products were
identified by mass spectrometry. Figure 4 shows a very general
diagram of a mass spectrometer system. The sample is introduced
into the source via an inlet system (gas or solid probe, reservoir,
12
-------
2
fl / E
Figure k- Mass spectrometer system (Adapted from Roboz, J. Introduction
to mass spectrometry. New York- Tn+.pTsci fince. 1963* P* 10)
-------
gas chromatograph) which also serves to volatilize the sample if
necessary. In the source, the sample gas is bombarded with elect-
rons to form ioris. Ttie positive ions are formed into a narrow
beam by passage through a slit and accelerated by a potential
difference into the mass analyzer. Here a magnetic field causes
the single ion beam from the source, which contains ions of many
different masses, to be resolved into many ion beams, each of
these beams containing ions of essentially one mass. Each of these
resolved ion beams is focused in turn on a detector and the re-
sponse amplified and recorded. A mass spectrum is simply a plot
of the relative abundance of an ion versus its mass-to-charge
ratio. The thing-that makes mass spectrometry a useful analytical
tool is that for a given energy of bombarding electrons the same
fragmentation pattern always results, providing a fingerprint of
the compound. Mass spectrometry which is much more sensitive than
infrared spectroscopy, can detect nanogram quantities of sample.
Since mass spectra of various components of a sample are additive,
however, only pure compounds or simple mixtures can be successfully
analyzed.
Most of the work was carried out using an AEI MS10 mass spectro-
meter (Figure 5) with a glass reservoir inlet system. Samples for
analysis by mass spectrometry were collected in the following way:
effluent which was not sent to the detector emerged from the GC
column through a two-way valve and was diverted into a U-tube,
filled with glass beads and cooled with liquid nitrogen, in order
to condense out everything but the helium carrier gas. The U-tube
and its cooling flask were then transferred to the inlet system
of the mass spectrometer, the helium was pumped out, and the cool-
ing flask was removed to allow the sample to expand into the
reservoir inlet. This inlet could not be heated, thus limiting
the analyses that could be performed to compounds boiling below
about 130 C, or "volatile" compounds. The MS10 mass analyzer
limited analyses to compounds with molecular weights less than 125.
During the last year of this grant period an AEI MS30 double beam
mass spectrometer was added to the laboratory (Figure 6), greatly
increasing the mass spectrometry capabilities. It included a gas
chromatograph inlet so that the sample could be injected onto a
chromatograph column and mass spectra of the eluted compounds
taken sequentially. The heated inlet and source allowed for anal-
ysis of compounds boiling up to 350 C. Liquid residues left after
incomplete combustion could thus be studied.
Quantitative Analysis
Infrared spectroscopy and gas chromatography are excellent tools
for quantitative analysis. For a wel1-resolved band in the infra-
red region, a plot of absorbance versus concentration for a series
1^
-------
-------
-------
of standards should yield a straight-line calibration curve. Like-
wise peak area (or peak height for a symmetrical, well-resolved
peak) is proportional to concentration in chromatography allowing
calibration curves of peak area (or height) versus concentration
to be made from a series of standards. Standards were prepared
in either a 20-liter glass carboy with mechanical stirrer or in a
SaranR bag containing a measured volume of air by injecting known
quantities of pure liquids or lecture-bottle gases.
A few compounds could not be quantitated by either infrared spec-
troscopy or gas chromatography. Oxides of nitrogen were quantita-
ted using the Saltzman method! 1 and sulfur dioxide was quantitated
by a recently-reported ultraviolet method J 2 Hydrogen cyanide and
ammonia were quantitated using Orion specific ion electrodes and
an Orion Model 801 digital pH meter.
Quantitative analysis of major products in some of the liquid
residues produced on combustion of many plastics was unsuccessful.
Weighing the combustion tube and bag or washing out the residue
with solvent proved exceedingly inaccurate; so the amount of
residue was determined by difference from volatile products.
Many tables of numbers are presented throughout the text and some
conment on their significance is in order. The analytical experi-
mental error due to such things as sample injection, standard pre-
paration, and peak measurement should be less than ±10 percent.
Another potential source of error was introduced by the method of
sample collection. Gaseous products came into contact with SaranR
(polyvinylidine chloride) collection bags and TygonR tubing. Both
materials were found to be inert to small concentrations of hydro-
carbon materials over the short period of time (less than two
hours) between sample collection and analysis. Anmonia, and pos-
sibly other nitrogen products, reacted slowly with SaranR. Use
of TeflonR bags was considered but was prohibitively expensive.
Condensation of higher-boiling products (>100 C) on the large sur-
face area of the bag was minimized by rapid sample analysis.
The rather random nature of the combustion process provided the
largest source of experimental deviation. The goal of the study,
largely accomplished, was to obtain values reproducible under our
combustion conditions to within ±25 percent. However, occasional
deviations of ±200 percent or more were observed, using identical
combustion conditions. Deviations this large were almost certainly
due to ignition of the plastic or its products in the heating zone.
Such deviations were largest for carbon dioxide, carbon monoxide,
and methane, the products of most complete combustion.
When comparing the data for carbon monoxide when burned with
varying amounts of air, it will be noted on some polymers, that
more of the compound is formed with increased amounts of air,
17
-------
contrary to what one usually expects, As pointed out earlier, we
calculated the amount of air necessary, when integrated over the
entire combustion run, to completely convert all the carbon to
carbon dioxide. The air supplies used were then less than, roughly
equal to, or more than that calculated. However, the amount calcu-
lated was a deficient supply because all the air passing through
the tube was not in the vicinity of the sample and thus available
for reaction. The reaction would only occur with reasonable effi-
ciency at the high end of the temperature range. This, coupled
with the possibility that some polymers are inclined to ignite in
the heating zone when the amount of air is increased, may result
in a greater reduction of the hydrocarbons to carbon without suff-
cient oxygen to convert it to carbon dioxide.
18
-------
POLi'VI'lYl CHLORIDE
Introduction
Polyvinyl chloride (PVC), of all the polymers, has been implicated
as causing the most serious solid waste disposal problem because
of its large volume usage in packaging and its release of hydro-
gen chloride gas on burning. The potential of this hydrogen chlo-
ride to corrode municipal incinerators has long been realized, and
has resulted in several states introducing legislatio^to ban PVC
packaging, for example, Michigan House Bill No. 5486. °
Polyvinyl chloride has the following structure:
H H
I I
C C
I I
H CI
Its chemical composition is 38.44 percent carbon, 4.84 percent
hydrogen and 56.73 percent chlorine. Low-boiling (<130 C) combus-
tion products of PVC have been studied prior to initiation of this
grant, and a summary of results is included.^ Previous workers con-
cerned with the combustion of vinyl chloride polymers examined
them primarily to determine the amounts of hydrogen chloride, car-
bon dioxide, carbon monoxide, and other major products given
off.15-18 Toxicity to rats of combustion products of the same
polymers used in this study has been reported.'9
Recently concern has been expressed over possible generation of
benzyl chloride on cutting PVC film with a hot wire. Since this
compound is quite toxic and might be liberated in any inefficient
PVC combustion (municipal incineration, home incineration, open
burning) its generation was specifically studied, particularly at
low temperatures (<350 C) as benzyl chloride decomposes at higher
temperatures.
Samples
Four PVC polymers and three formulations of three of the polymers
were analyzed under previous grants and are listed as follows:^
polymers A, B, and C, polyvinyl chloride homopolymers; polymer D,
an 85:15 copolymer of vinyl chloride and vinyl acetate; plastic
E, a common wire insulation formulation containing 57 percent
polymer C; .plastic F, a floor tile formulation containing 35 per-
cent copolymer D; and plastic G, a wire insulation formulation
19
-------
containing 51 percent polymer B, Five comnercial meat-wrap films
(A,B,C,D,E) were tested for benzyl chloride production under the
present grant.
Res u 1 ts
Thermal analysis. Differential thermal analysis of polymer C
heated at 10 C/min in excess air (Figure 7) shows a dehydrochlo-
rination endotherm at about 310 C, partly obscured by much larger
exothermic peaks which begin at 260 C and continue to 600 C, at
which temperature the sample is completely combusted. Thermogravi-
metric analysis of the same polymer heated at 3 C/min in air
(Figure 8) shows the weight loss appears to take place in five
stages. First is a rapid 60-percent loss up to 280 C corresponding
to the DTA endotherm; second, a decreasing rate of loss up to 350 C;
third, a slow constant rate of loss to 430 C; fourth, a more rapid
rate of loss to 510 C; and finally, a faster rate of loss, for the
remainder of the sample. These temperature ranges are used later
(Table 4) in describing the change in composition of combustion
products with temperature. DTA and TGA records for the other
polymers and plastics showed similar steps with ±20-C variations
in the temperatures. Thermal analysis was not carried out on the
film samples.
Qualitative analysis. This study was primarily concerned with
"volatile" products of combustion, or those compounds boiling up
to about 150 C. Analysis of the combustion gas by infrared spect-
roscopy showed absorption bands of hydrogen chloride, carbon di-
oxide, carbon monoxide, and benzene, along with other bands in the
3000 cm-1 region indicating the presence of compounds containing
C-H groups. Identification of these other compounds was carried
out by separating them from the mixture by gas chromatography and
collecting them in individual quantities for identification by
mass spectrometry using an AEI MS10. A six-foot-long, 1/4-in.-
diaineter column of Porapak Q was used to separate compounds boil-
ing at temperatures less than 0 C. Higher-boiling compounds were
separated on a 12-foot-long, 1/4-in.-diameter column of 5 percent
Squalane on Chromosorb P. Columns were temperature-programmed at
10 C/min to 110 C after an initial period of 5 to 15 minutes at
25 C.
Shown in Figures 9 and 10 are the 51 chromatographic peaks eluted
on these two columns representing a minimum of 59 volatile pro-
ducts of combustion (some peaks represent more than one compound)
including the three identified by direct infrared analysis. Of
this minimum figure of 59, 52 compounds have been identified
including carbon dioxide, carbon monoxide, hydrogen chloride and
the 49 listed in Table 2.
20
-------
T(°C)
Figure 7- Differential thermal analysis record of polymer C
heated at 10 C/min in air
-------
T (°C)
Figure 8. Thermogravimetric analysis record of polymer C heated at J c/min in air
-------
Figure 9. Chromatogram of low-boiling combustion products of polymer A on a Porapak Q column
-------
Figure 10. Chromatogram of combustion products of polymer A on a column of 5 percent squalane on ChromosorbP
-------
TF-LuZ 2
IDENTIFICATION Of POLYVIN't?. CHLORIDE CHROMATOGRAM PEAKS
Peak no. Identification
1 Methane
2 Ethylene
3 Ethane
4 Propylene
5 Propane, Methyl chloride
6 Vinyl chloride
7 1-Butene, Isobutane, Butadiene
8 Butane
9,10 trans-2-Butene, cis-2-Butene
11 3-Methyl-l-butene
12 Isopentane, 1,4-Pentadiene
13 1-Pentene
14 Pentane
15,16 trans-2-Pentene, cis-2-Pentene
17 2-Methyl-2-butene
18 cis or trans-1,3-Pentadiene
19 cis or trans-2-Penten-4-yne
20 Cyclopentene
21 Cyclopentane
22 2-Methylpentane
23 1-Hexene, 3-Methylpentane
24 Hexane
25 2-Hexene
26 Methylcyclopentane
27 1-Methylcyclopentene
28 Benzene
29 Cyclohexane
30 1-Heptene
31 1,4-Dimethylcyclopentene*
32 Heptane
33 Unidentified
34 3-Ethylcyclopentene
35 Methylcyclohexane
36 Ethylcyclopentane
37 1,2-Dimethylcyclopentene
38 1-or 4-Ethylcyclopentene
39 1-Methyl eyelonexene
40 Toluene
41,42 Unidentified
43 Octane
44-47 Unidentified
48 Ethyl benzene
49-51 o-, m-, p-Xylene
*Tentative
25
-------
Quantitative analysis. All quantitative work was done on the 22
compounds present in the greatest quantities. As a uniform con-
dition for comparing polymers and formulations, the following com-
bustion conditions were adopted: The air supply was 60 cc/min
which, when integrated over the entire run, would result in about
twice the amount of oxygen necessary to convert all the carbon to
carbon dioxide. The plastics were heated (after an initial heat-
ing from room temperature) from 200 C to 600 C at a rate of 3
C/mi n.
Hydrogen chloride was quantitated by infrared spectroscopy, acid-
base titration, and silver nitrate titration, and in the case of
all the hotnopolymers amounted to about 580 milliqrams Der qram.
Thus, HC1 formed nearly quantitatively from the chlorine atoms of
PVC. . .
A comparison of the combustion products of the three plastics with
the combustion products of their polymers is given on Table 3.
Except for aromatic compounds, the hydrocarbons in plastics G and
E have all increased in quantity over polymers B and C by factors
ranging from 1.3 to 8 times. Likewise, the amount of vinyl chlo-
ride appears to be about 5 times as great in these samples. Most
of the increases are attributable to the breakdown of the phtha-
late plasticizer, which forms a series of hydrocarbons similar to
those produced by PVC. The plasticizer, either dioctyl phthalate
of diisodecyl phthalate, cannot be directly responsible for the
apparent increase in vinyl chloride.
Different results are noted for plastic F, the floor tile formu-
lation made from copolymer D. This product contains about one
third copolymer D, but the hydrocarbon products generated, espe-
cially saturated aliphatics and benzene, are considerably less
than one third the amount generated by copolymer D. Likewise, the
amounts of acetic acid and HC1 are less than the 33 percent
expected. This product contains about 70 percent inert material,
such as asbestos and calcium carbonate, which may play a part in
inhibiting breakdown of the polymer and production of hydrocarbons.
The variations in quantities of combustion products of a represent-
ative PVC homopolymer as a function of temperature are shown in
Table 4. Products were collected in five fractions, selected on
the basis of TGA curve inflections, during a single heating run.
During the first temperature fraction almost 80 percent of the
benzene is formed along with a small amount of toluene and some
unsaturated hydrocarbons. Production of hydrogen chloride as
determined.by analysis of the TGA effluent roughly parallels that
of benzene. In the second fraction (280-350 C), carbon dioxide
and carbon monoxide appear, toluene continues to increase, but
hydrogen chloride and benzene are already decreasing and continue
to decrease through the higher ranges. Hydrogen chloride is present
26
-------
TABLE 3
COMPARISON OF COMBUSTION PRODUCTS OF THE PLASTICS WITH THE COMBUSTION PRODUCTS OF THEIR POLYMERS
Compound
Polymer
B
PI asti c
G
Polymer Plastic
C E
Copolymer
D
PI asti c
F
Hydrochloric acid*
583.
273.
584.
333.
500.
73.
Acetic acid
96.
20.
Carbon dioxide
729.
616.
730.
1182.
923.
456.
Carbon monoxide
442.
67.
403.
90.
292.
31 .
Me thane
4.6
6.6
5.8
6.8
4.4
0. 30
Ethylene
0.58
2.3
0.33
2.0
0.60
:. :2
Ethane
2.2
3.0
2.5
2.9
2.3
v. 1 J
Propylene
0.47
2.0
0.56
1.4
0.56
0.11
Propane
0.84
1.7
1.1
1,4
0.88
O.K.
Vinyl chloride
0.60
3.3
0.52
2.6
0.72
0.3:
1-Butene
0.18
1.1
0.28
0.58
0.22
O.Jo
Butane
0.28
1.1
0.39
0.74
0.29
0.05
Isopentane
0.02
0.15
0.02
0.04
0.02
0.01
1-Pentene
0.06
0.35
0.11
0.15
0.09
0.01
Pentane
0.16
0.58
0.27
0.38
0.21
0.02
Cyclopentene
0.05
0.14
0.58
0.07
0.05
0.004
Cyclopentane
0.05
0.16
0.07
0.09
0.06
0.003
1-Hexene
0.05
0.24
0.09
0.18
0.08
0.01
Hexane
0.12
0.49
0.25
0.35
0.17
0.01
Methylcyclopentane
0.14
0.14
0.07
0.09
0.05
Benzene
36. \
10.
29.
11.
28.
0.86
Toluene
1.3 \
0.94
1.1
1.0
0.96
0.04
Residue (inorganic)
\
159.
61 .
709.
*The quantity
of each combustion
>
\product is
reported in
milligrams per gram of sample.
-------
TABLE 4
VARIATION OF COMBUSTION PRODUCTS OF POLYMER A WITH TEMPERATURE
Compound
25-
280-
350-
430-
510-
280 C
350 C
430 C
510 C
580 C
Carbon dioxide*
-
9.7
181 .
244.
237.
Carbon monoxide
20.
46.
151 .
181 .
Methane
0.20
1.3
1 .8
0.31
Ethylene
' 0.04
0.33
0.39
Ethane
0.12
0.94
0.41
Propylene
0.06
0.11
0.31
Propane
0.08
0.44
0.11
Vinyl chloride
0.04
0.25
0.17
0.02
1-Butene
0.02
0.04
0.08
Butane
0.03
0.20
0.02
Isopentane
0.005
0.001
1-Pentene
0.01
0.03
Pentane
0.01
0.08
0.01
Cyclopentene
0.02
0.01
Cyclopentane
0.01
0.02
Hexane
0.01
0.05
0.01
Methylclopentane
0.02
Benzene
24.
6.6
0.35
0.16
Toluene
0.12
0.18
0.55
0.03
0.01
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
only in trace amounts after 300 C. Additionally, in the second
fraction the maximum amount of vinyl chloride is formed. As
expected, carbon dioxide and carbon monoxide reached their maxima
at higher temperatures. Most straight-chain aliphatics reached
their maxima in the third step and were present in the last step
in very small quantities. Olefins began to form in the first step,
reached their maxima in the third, and did not appear in the last
two fractions. More detailed information concerning variation of
combustion products with heating rate and air supply may be found
in Reference 14.
Benzyl chloride analysis. Analysis for benzyl chloride was con-
ducted independently of analysis for other combustion products
and only on five commercial meat-wrap films. Heating procedure
differed from the standard procedure. The 0.25-g samples were
heated at the bottom of a 250-ml round-bottom flask enclosed in a
heating mantle. In close proximity to the sample in the flask
was a thermocouple which was used both to monitor and control the
temperature. In some tests the sample and controlled thermocouple
were placed at the bottom of a vial, and covered with about one
cm of MallcosorbR to remove the hydrogen chloride gas. The vial
was placed on the bottom of the heating flask.
The stopper in the flask contained two short glass tubes, one
capped by a septum through which the evolved gases in the flask
were sampled, using a 10-cc gas syringe. The other tube was
connected to a small SaranR plastic bag, which acted as an expan-
sion chamber as the flask was heated. When the plastic reached
the desired temperature (and periodically thereafter), one to
three cc of gas were removed from the flask with a gas syringe,
and transferred to either the gas chromatograph or the GC-MS30.
A six-foot-long, 1/4-in.-diameter column of 10 percent Carbowax
20 M on Chromosorb P was used to separate benzyl chloride from
other products. At 100 C the net retention time of benzyl chloride
was 36 minutes. This peak was positively identified as benzyl
chloride by GC-MS.
Films A and B were heated to 200 C and 337 C (using the heating
arrangement described) and held at these temperatures until past
the emission of benzyl chloride. The heating cycle was begun at
time zero and 30 to 60 minutes were required for the flask to
attain the desired temperatures. The data, plotted in Figures 11
and 12, show 0.1 to 5 milligrams benzyl chloride per gram of
sample. These runs were made without the use of MallcosorbR,
except in one run as noted in Figure 11. Because the increase in
benzyl chloride at the start of the runs corresponds with the
increase in temperature, it is suspected that the amount of benzyl
chloride is proportional to the temperature, with only a secondary
dependence on heating rate, but this hypothesis should be veri-
29
-------
TIME AFTER START OF HEAT (min)
Figure 11. Benzyl chloride production by two meat-wrap
films (A and B) heated slowly to 200 C
30
-------
films (A and B) heated slowly to 337 C
31
-------
fied, especially for very fast heating,
The flask was then preheated to 200 C and the sample was dropped
into the bottom of the flask which was immediately re-stoppered.
Maintaining the flask at 200 C, samples were taken at periodic
intervals, resulting ir; 0,01 to 5 milligrams benzyl chloride per
gram of sample as showr. graphically in Figure 13. Similarlly, the
data in Figure 14, showing 0.05 to 10 milligrams benzyl chloride
per gram of sample, were obtained by preheating the flask to 337 C,
and maintaining it at that temperature after inserting the sample.
Although these data confirm the conclusion that the maximum amount
of benzyl chloride observed was primarily proportional to the
temperature, (although testing only two temperatures), it must be
pointed out that the faster heating rate is still considerably
slower than encountered in hot-wire cutting.
Film C was identical to Film B and Film D was identical to Film A.
Film E had an additive and produced slightly less benzyl chloride
than Film A. The amount of benzyl chloride produced depends on
the batch of plastic, with about one batch in 20 producing the
larger amount. The reason for these batch differences has not
been determined,
Pnosgene analysis. Phosgene has occasionally been reported as a
combustion product of PVC, although it has not been detected in
any of our work. A report that this compound is produced only at
high temperatures led us to propose a 1aser-pyrolysis study. A
preliminary test with a carbon dioxide laser available to us at
the University of Rochester was carried out but was unsuccessful
because of insufficient power. It is hoped that the experiment
can be repeated later with a higher-powered laser.
Pi scussion
Quantitative release of hydrogen chloride on thermal degradation
of PVC presents a serious problem'in its disposal. Incinerator
stack gases may be wet-scrubbed, effectively removing hydrogen
chloride. However, hydrogen chloride is still implicated in incin-
erator corrosion processes, even though these processes have not
been fully elucidated.8 The fact that the calcium carbonate-filled
floor tile sample showed less hydrogen chloride than expected
indicated there may be some promise in the method of spreading some
such chemical on material to be incinerated in order to react with
or absorb hydrogen chloride.
Our data fits rather well with a degradation mechanism proposed by
Madorsky5 involving rupture of the C-Cl bond followed by abstrac-
tion of an adjacent hydrogen atom, forming a double bond. The
32
-------
TIME AFTER INSERTING SAMPLE (min)
Figure 13. Benzyl chloride production by two meat-wrap
films (A and B) heated at 200 C
33
-------
>
N
Z
LU
GQ
10
-I
I0"2
FLASK PREHEATED AND
MAINTAINED AT 337°C
x-FILM
A
60 120 180
TIME AFTER INSERTIN6 SAMPLE (min)
Figure lU. Benzyl chloride production by two meat-wrap
films (A and B) heated at 337 C
34
-------
chain reaction involves successive remoyal of chlorine atoms in
the g position to the first double bond forming a chain with con-
jugated double bonds. Benzene is thermodynamically the most stable
molecule that could be formed from the resulting chain containing
conjugated double bonds. The fact that benzene forms during the
same temperature range as hydrogen chloride and before other hydro-
carbons is significant. Toluene and benzyl chloride could be
formed by occasional skips in the randomly conjugated double-Dona
formation.
35
-------
POLYSULFONE
Introduction
Polysulfone (Figure 15) is a specialty thermoplastic presently
produced in comparatively small quantities for electrical, auto-
motive, and appliance parts.^I Primary interest in its combustion
products was to account for its approximately 7.6 percent sulfur.
Samples
Two polysulfone samples in commercial pellet form were studied.
Results
Thermal analysis. Differential thermal analysis of polysulfone
was carried out in both helium and air atmospheres using a 10-t/min
heating rate (Figure 16)*. The plastic was liquid by 350 C, but
no melting endotherms are evident. The only significant features
of either curve are two exothermic peaks at 535 and 645 C in air,
both corresponding to temperatures where production of volatile and
nonvolatile products occurs very rapidly. Polysulfone thermogravi-
metric analysis records were also obtained in both helium and air
atmospheres (Figure 17). The heating rate was 5 C/min. These
records show a 62-percent weight loss between 460 and 520 C in
helium, while in air a 55-percent weight loss occurs more rapidly
between 460 and 500 C. The remaining 45 percent of the weight in
air is lost much more gradually above 510 C. Davis has shown that
crosslinking is an important factor in polysulfone degradation
although chain scission is also involved.
Qualitative analysis. Absorption bands characteristic of carbon
dioxide, carbon monoxide, carbonyl sulfide, and sulfur dioxide
appear in the infrared spectrum of polysulfone combustion products
(Figure 18). The spectrum was obtained using a 10-cm path-length
gas cell. Hydrogen sulfide has been reported as a combustion
product of polysulfone,23 but it would not be observed in small
quantities with either the infrared spectrophotometer or the hydro-
gen flame gas chromatograph detector used in this study. Methane,
ethylene, and ethane were separated by gas chromatography on a
six-foot-long, 1/4-in.-diameter Porapak Q column. Benzene, toluene,
ethylbenzene, and styrene were separated on a 12-foot-long, 1/4-in.-
diameter column of 5 percent Squalan'e on Chromosorb P. Those were
the only gaseous products identified. Thirty to fifty percent of
the polysulfone combusted formed liquid residue which condensed in
the combustion tube or sample bag. Components of this liquid
residue were separated on a 13-in.-long, 1/4-in.-diameter low K'
Durapak column (Figure 19) and have been characterized or identi-
36
-------
-©¦r-o-oj-o
Un i U
^4
-3
n
Figure 15. Bisphenol A-polysulfone
-------
AT
Ex
En
V
J I I I I I
100 200 300 400 500 600
T(°C)
Figure 16. Differential thermal analysis records of polysulfone heated at 10 C/min
in helium (broken curve) and air (smooth curve)
-------
T (°C)
Figure 17- Tliermograviraetric analysis records of polysulfone neated a~ ^ c/min
in helium (broken curve) and air (smooth curve)
-------
FREQUENCY (CM1)
WAVELENGTH (MICRONS)
igure IS- -A portion of an infrared spec ^rum of polysulfone
combustion prciucts (10-cm path-lengtn
-------
TIME ¦
Figure 19. Chromatogram of polysuifone liquid residue on a low K' Z-^rapak column
-------
fied (Table 5). Major components of the liquid residue are phenol,
cresol, and phenyl-p-tolyl ether. Isomers were difficult to dis-
tinguish as many peaks were due to multicomponent mixtures of
products.
Quantitative analysis. Polysulfone samples for quantitative ana-
lysis were heated rapidly from room temperature to 350 C and then
programmed at b to 50 C/min to either 800 or 1000 C. Table 6
shows quantities of eight gaseous products produced by one of the
polysulfone samples at five combustion conditions with the 800 C
temperature maximum. Carbon dioxide, carbon monoxide, and sulfur
dioxide were quantitated by infrared spectroscopy using a 10-cm
path-length gas cell. Sulfur dioxide, however, tended to dissolve
in any moisture present in the cell and to be adsorbed onto metal
parts, thus leading to erratic results. It was quantitated again
by ultraviolet spectrophotometry^ at each of the combustion
conditions. The two polysulfone samples tested had 7.62 and 7.56
percent sulfur, and so about 76 mg sulfur had to be accounted for
in each gram of plastic. The hydrocarbons were quantitated by
gas chromotography on the same columns used for their separation.
Quantities in Table 6 show no discernable trends.
Quantities of combustion products of polysulfone using the 1000 C
temperature maximum are shown in Table 7. Results for the two
samples were averaged and were often nearly identical, as indicated
by the calculated deviations which include differences between the
two samples as well as experimental error and the randomness of
the combustion process. In all cases over 85 percent of the sul-
fur of polysulfone is accounted for as sulfur dioxide. Volatile
products identified account for 50 to 60 percent of the plastic
combusted, the remaining 40 to 50 percent being liquid residue.
Quantities of compounds produced during several temperature ranges
chosen on the basis of the TGA curve are shown in Table 8. Carbon
dioxide and carbon monoxide are absent below 450 C and only a
small amount of the latter is formed by 490 C. Sulfur dioxide is
formed primarily between 490 and 550 C, where carbonyl sulfide is
also found. This compound was probably too dilute to be detected
in samples taken over the entire temperature range of 0 to 1000 C.
Aliphatic hydrocarbons peak in the third temperature range while
aromatic hydrocarbons peak in the fourth. Again, over 90 percent
of the sulfur and about 60 percent of the plastic is accounted for.
Discussion
Thermal degradation of polysulfone seems to proceed by chain
scission so that most of the sulfur is released as sulfur dioxide.
Relatively small amounts of carbonyl sulfide and possibly hydrogen
sulfide also formed. The most significant hydrocarbon products,
toxicologically, were found in the liquid residue.
42
-------
TABLE 5
IDENTIFICATION OF POLYSULFONE RESIDUE CHROMATOGRAM PEAKS
Peak
Molecular
Number
Weight
Identi fi cation
1
92
Toluene
2
106
Ethylbenzene (or Xylene)
3
104
Styrene
4
120
Methyl ethylbenzene*
126
Chiorotoluene?
5
106
Benzaldehyde
118
Benzofuran
116
Indene
6
132
Methylbenzofuran
130
Methylindene
120
Trimethyl benzene*
7
128
Naphthalene
142
Methyl naphthalene
8
94
Phenolt
154
Biphenyl
9
170
Diphenyl ether (or
Phenyl phenol)
10
108
Cresol
11
168
Dibenzofuran
184
Phenyl-p-tolyl ethert
156
Dimethylnaphthalene
122
Ethyl phenol
12
198
aryl-Ethylphenyl phenyl
ether*
13
212
2-Hydroxyphenyl-2-phenyl
propane
14
226
Unidentified
15
224
Unidentified
* Or isomer
+ Major component
43
-------
TABLE. 6
COMBUSTION PRODUCTS OF POLYSULFONE AT SEVERAL COMBUSTION CONDITIONS
(800 C miximum)
Air flow, cc/min
65
98
98
200
450
Oxygen flow, cc/min
—
--
20
—
—
Heating Rate, C/min
5
5
5
5
50
Carbon dioxide*
1331
1044
1148
1661
1488
Carbon monoxide
298
363
253
437
265
Sulfur dioxide-IR
204
152
132
169
179
Sulfur dioxide-UV
150
149
149
152
150
Methane
19.0
18.0
32.6
14.8
16.2
Ethyl 6'".e
0.77
0.53
0. 73
0.41
0.67
Ethane
0.68
0.86
0.72
0.67
0.74
Benzer.-
5.23
5.71
4.34
4.45
6.47
7 ol uetic
2.19
2.52
1.55
T .96
3. 71
Sulfur accounted for (UV) 98.7
98.0
98.0
100.
98.7
--Pi as -" c accounted for
67.
62.
61.
82.
70.
^Residue (ay difference)
33.
38.
39.
18.
30.
x7ne quantity of each combustion product is reported in milligrams per gram of sample.
-------
TABLE 7
COMBUSTION PRODUCTS OF POLYSULFONE AT SEVERAL COMBUSTION CONDITIONS
(1000 C maximum)
Air flow, cc/min 100 100 100
Oxygen flow, cc/min 0 40 0
Heating rate, C/min 5 5 50
Carbon dioxide*
1072
+
125
861
+
47
"7^i4
+
82
Carbon monoxide
250
+
4
261
+
0^
87
+
6
Sulfur dioxide
145
+
1
1
8
131
+
4
Methane
28.4
+
0 a.
9.68
+
0.20
23.5
4-
0.10
Ethylene
O-T-7--S-
uT02
0.63
+
0.16
1.17
+
0.06
Etnane
0.72
±
0.03
0.60
+
0.18
0.60
+
0.03
Benzene
5.39
±
0.03
4.76
±
0.01
8.69
+
1.57
Toluene
2.59
±
0.01
2.16
j.
0.25
3.61
+
1.17
Ethyl benzene
0.60
±
0.03
0.53
+
0.06
0.67
+
0.13
Styrene
0.12
±
0.01
0.12
0.04
0.20
+
0.08
xSulfur accounted for
95.6
91 .6
86.0
/.Plastic accounted for
58
50
59
^Residue (by difference)
42
50
41
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
TABLE 8
POLYSULFONE COMBUSTION PRODUCTS DURING SEVERAL TEMPERATURE RANGES
Compound
<450 C
450 -
490 C
490 -
550 C
>550 C
Total
Carbon dioxide*
_
_
31
1104
1135
Carbon monoxide
2.28
35
213
250
Sulfur dioxide
6.75
29.3
93
11.3
140
Caroonyl sulfide
0.52
0.52
Methane
0.38
2.86
10.2
6.59
20.0
Ethylene •
0.01
0.10
0.39
0.33
0.83
Ethane
0.01
0.06
0.33
0.14
0.54
Benzene
0.10
1 .64
3.82
5.56
Toluene
0.01
0.02
0.54
1.18
1 .75
Ethylbenzene
0.01
0.04
0.37
0.42
Styrene
- - - -
— — — —
— — — —
%Su1fur accounted for 92.6
%Plastic accounted for 60
^Residue (by difference) 40
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
POLYURETHANE
Introduction
Polyurethane applications include furniture, mattresses and pil-
lows, and automobile interiors. Interest in combustion products
of polyurethanes stems from three considerations: first, although
not produced in as large quantities as polyethylene or polyvinyl
chloride, polyurethanes can contribute significantly to the overall
incineration problem because of their use for articles which are
so bulky they are difficult to bum efficiently. Second, such
articles tend to smolder in an accidental fire producing fumes of
unknown toxicity. Third, the isocyanate materials used in poly-
urethane production can cause severe sensitization reactions in
some people. Because of isocyanate contact during polyurethane
production,24 most studies of the toxicity of polyurethane or its
combustion products have been directed toward analysis of any
isocyanate present. The current study has been directed toward
the lower-boiling combustion products and has not yet been extended
to include isocyanates.
Six commercial polyurethane samples were analyzed. Four were foam
samples of the type used for automobile seats and carpet backing,
and are listed as follows:
In addition, two samples in pellet form were obtained: Sample E,
an aromatic methylene diisocyanate compound containing 4.15 per-
cent nitrogen; and Sample F, an H-^-aliphatic methylene di iso-
cyanate compound containing 4.07 percent nitrogen. Both are of
the polyester type.
Thermal analysis. The differential thermal analysis record of one
of the foam samples (Figure 20) shows an exothermic doublet between
290 and 375 C, peaking at 310 C. No endotherms are evident. The
thermogravimetric analysis record of a foam sample (Figure 21)
shows a two-step degradation, the first and largest step beginning
at about 260 C and ending at about 310 C, during which 77 percent
of the weight is lost. The remaining step is very gradual and
the plastic is completely combusted by 575 C. Thermogravimetric
analysis of the aromatic pellet sample (Figure 22) shows that the
same two-step degradation occurs but at higher temperatures.
Samples
Foam A
Foam B
Foam C
Foam D
4.74 % N
4.62 % N
5.57 % N
4.26 % N
Results
47
-------
AT
Ex
-p-
OO
En
1
i
100 200
300
T
400
(°C)
500
600
Figure 20. Differential thermal analysis record of a polyurethane foam heated at 10 c/min in air
-------
T (°C)
Figure 21. Thermogravimetvie analysis record of a polyurethane foam i:ea-ed at 10 C/rair. ir. air
-------
100
75
e>
2
UJ
* 50
i-
$
25 -
100
200
Figure 22. Thermogravimetric analys
300 400 500 600
T(°C)
record of polyurethane Sample E heated at 10 C/min in air
-------
Sample E does riot begin to lose weigh I until 300 C and then loses
85 percent of its weight by 500 C. Tl-/} remainder of the weight is
lost more gradually up to 625 C.
Qualitative analysis. Qualitative analysis of combustion gas from
the polyurethane foam samples by infrjred spectroscopy using a
10-cm path-length gas cell revealed binds due to carbon dioxide,
carbon monoxide, methanol, and acetal lehyde. The pellet samples
did not produce the latter two compounds. For the foam samples,
C]-through (^-hydrocarbons, hydrogen vyanide, and cyanogen were
separated using a six-foot-long, 1/4-jin .-diameter Porapak Q col-
umn while methanol, acetaldehyde, pre )ionaldehyde, and acetone
were separated on a six-foot-long, 1/1-in-diameter column of 10
percent Carbowax 20 M on Chromosorb F. Compounds were identified
by mass spectrometry. The pellet samples were analyzed later in
the study and a six-foot-long, 1/4-irt.-diameter column of Durapak
n-Octane/Porasi1 C was used for separation of products because its
resolution of hydrocarbon isomers was! superior to that of Porapak
Q. C-j-through Cg-hydrocarbons could be janalyzed using the Durapak
column, but not oxygen-containing compounds such as methanol and
acetone. Preliminary GC-MS analysis of(liquid residues formed on
combustion of polyurethanes has shown that toluidine and either
aniline or methyl pyridine are significant components. Halogenated
compounds were observed but not positively identified. The pre-
sent study of polyurethane combustion products is incomplete and
many volatile and nonvolatile products remain to be identified.
Other studies 25,26 have identified more nitrogen-containing pro-
ducts sucJi as acetonitrile, acrylonitrile, pyridine, and phenyliso-
cyanate. Products have not been quantitated, however, and parts
of these studies were done under pyrolysis conditions.
Quantitative analysis. Products, except hydrogen cyanide and car-
bon oxides, were quantitated on the chromatograph columns used
for their separation. Carbon dioxide and carbon monoxide were
quantitated by infrared spectroscopy. Cyanide ion was quantitated
using an Orion specific ion electrode and an Orion 801 digital pH
meter. This electrode measures only cyanide ion and not the
equally toxic organic cyanide compounds (nitriles). Ammonia was
quantitated using an Orion ammonia-specific electrode. Oxides of
nitrogen from several foam samples were quantitated by the
Saltzman method but were found to be in the microgram per gram
range. Tables 9 and 10 show milligrams of identified combustion
products per grain of foam sample when four foams were combusted
under several different conditions. Cyanide ion, calculated as
hydrogen cyanide, ranges from 7 to 46 milligrams per gram of sam-
ple and accounts for 9 to 53 percent of the nitrogen of the plastic
combusted. A significant decrease in cyanide ion is noted at the
faster heating rate, indicating it is either broken down or not
formed as efficiently at higher temperatures. Other products, and
the hydrocarbons in particular, increase considerably at the
51
-------
TABLE 9
COMBUSTION PRODUCTS OF POLYURETHANE FOAMS A AND B
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Foam A
Carbon dioxide*
712.
661.
533.
Carbon monoxide
193.
207.
169.
Cyanide ion (as HCN)
46.2
34.6
19.3
Ammoni a
1.60
4.42
0.24
Methane
1.73
2.02
21.2
Ethylene
1.95
3.55
16.9
Ethane
0.18
0.60
3.9
Propylene
2.31
10.55
67.3
Propane
0.22
0.60
7.3
Methanol
13.2
9.3
.6.6
Acetaldehyde
17.7
20.5
10.2
Propionaldehyde
7.2
13.7
7.8
Acetone
13.6
14.0
12.5
^Nitrogen accounted for
53.3
45.4
21.1
%P1astic accounted for
38.3
38.2
42.8
Foam B
0
Carbon dioxide
657.
568.
521.
Carbon monoxide
116.
159.
211.
Cyanide ion (as HCN)
40.5
32.7
19.3
Amnoni a
N.A.+
N. A.
N.A.
Methane
1.86
2.23
21.6
Ethylene
2.32
4.01
17.6
Ethane
0.33
0.73
5.0
Propylene
2.84
10.40
42.8
Propane
0.30
0.40
10.7
Methanol
13.6
4.7
14.2
Acetaldehyde
18.2
20.9
28.5
Propionaldehyde
7.5
17.7
21.9
Acetone
15.2
15.6
34.3
% Nitrogen accounted for
45.5
36.7
21.6
%Plastic accounted for
33.1
33.2
44.8
*The quantity of each combustion product is reported in
milligrams per gram of sample.
tNot analyzed.
52
-------
TABLE 10
COMBUSTION PRODUCTS OF POLYURETHANE FOAMS C AND D
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Foam C
Carbon dioxide*
591.
836.
480.
Carbon monoxide
165.
235.
241.
Cyanide ion (as HCN)
34.7
32.5
11 .6
Ammoni a
0.23
0.09
0.01
Methane
2.40
2.41
27.3
Ethylene
1.57
3.08
15.5
Ethane
0.27
0.32
3.5
Propylene
3.56
10.60
61.1
Propane
0.37
0.68
8.3
Methanol
26.4
19.7
26.0
Acetaldehyde
27.1 .
32.5
53.9
Propionaldehyde
2.4
3.1
16.2
Acetone
12.4
11.1
39.8
%Nitrogen accounted for
32.3
30.3
10.8
%Plastic accounted for
34.3
44.5
49.8
Foam D
Carbon dioxide
570.
486.
395.
Carbon monoxide
295.
275.
311.
Cyanide ion (as HCN)
23.1
27.0
7.3
Ammonia
N.A.f
N.A.
N.A.
Methane
2.65
1.42
36.0
Ethylene
3.26
2.11
24.7
Ethane
0.38
0.25
6.5
Propylene
5.80
4.76
79.7
Propane
0.84
0.54
17.2
Methanol
33.5
20.7
31.2
Acetaldehyde
26.8
35.4
70.2
Propionaldehyde
3.2
4.5
18.8
Acetone
15.5
15.0
45.6
%Nitrogen accounted for
28.1
32.9
8.9
%Plastic accounted for
39.6
36.3
57.8
*The quantity of each combustion product is reported in
milligrams per gram of sample.
tNot analyzed.
-------
TABLE 11
COMBUSTION .PRODUCTS OF POLYURETHANE SAMPLE E
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Carbon dioxide*
361.
482.
267.
Carbon monoxide
110.
191.
49.
Cyanide ion (as HCN)
34.1
26.9
7.28
Ammonia
0.09
0.01
0.05
Methane
4.78
3.17
3.83
Ethylene
1.15
1.13
5.74
Ethane
1.82
1.61
1.28
Propylene
3.16
2.89
4.63
Propane
1.19
1.36
0.90
1-Butene
1.39
1.42
2.42
Butane
1.00
1.02
0.74
trans-2-Butene
7.80
8.08
14.2
cis-2-Butene
0.29
0.28
0.45
Pentane
1.10
1.19
0.37
1,3-Pentadiene?
11.4
12.2
7.71
1-Hexene
0.04
0.04
0.04
2-Hexene
0.17
0.22
0.19
%P1asti c accounted for
21.5
o 27.5
14.3
*The quantity of each combustion product is reported in
milligrams per gram of sample.
faster heating rate, perhaps because the polymer breaks down at
such a low temperature that carbon dioxide and carbon monoxide are
not efficiently formed. In most cases the volatile products quan-
titated account for 30 to 50 percent of the foam combusted. Quan-
tities of combustion products of Samples E and F at three combus-
tion conditions are listed in Tables 11 and 12. Cyanide ion again
decreases at the faster heating rate. Samples D and F appear to
generate more cyanide ion at the 40r-cc/minute oxygen flow indicat-
ing that the trend of polyurethane cyanide ion generation cannot
be safely generalized. Amounts of hydrocarbons generated are in
the 1-to 10-milligram per gram range, and increase somewhat at
the faster heating rate. The pentadiene identification has not
been confirmed. Only 15 to 30 percent of these samples can be
accounted for by the volatile products thus far identified.
Although a considerable amount of liquid residue formed on combus-
54
-------
TABLE 12
COMBUSTION PRODUCTS OF POLYURETHANE SAMPLE F
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Carbon dioxide*
380.
690.
259.
Carbon monoxide
164.
223.
98.
Cyanide ion (as HCN)
15.1
40.5
6.47
Ammoni a
0.09
2.03
0.09
Methane
2.89
1.50
3.37
Ethylene
6.68
4.20
4.12
Ethane
1.46
0.66
0.44
Propylene
10.6
3.65
10.0
Propane
0.89
0.78
0.45
1-Butene
1 .78
1.73
2.86
Butane
0.45
0.66
0.19
trans-2-Butene
4.09
2.70
11.2
cis-2-Butene
0.27
0.12
0.24
1-Pentene
0.45
0.18
0.69
Pentane
0.31
0.12
N.A.+
1,3-Pentadiene?
5.56
2.52
5.74
1-Hexene
0.17
0.12
0.17
2-Hexene
0.45
0.38
0.42
%Plastic accounted for
31.3
34.6
. 15.8
*The quantity of each combustion product is reported in
milligrams per gram of sample.
+Not analyzed.
tion of these samples, the amount of residue cannot be determined
by difference because it is suspected that many more volatile pro-
ducts remain to be identified and quantitated.
Discussion
Studies to date indicate that combustion of polyurethanes may re-
sult in a wide variety of hydrocarbon and nitrogen-containing
products. Carbon monoxide and cyanide are the only acutely toxic
compounds identified. Neither would be expected to form in signi-
ficant quantities during incineration, but large quantities could
form during open burning or an accidental fire. Since carbon
monoxide disrupts the bloods ability to carry oxygen and cyanide
disrupts the cell's ability to use oxygen, there is a toxicolog-
55
-------
ical synergism between these two compounds making their combina-
tion particularly hazardous.
56
-------
POLYIMIDE
Introducti on
Polyimide is one of the most temperature-resistant plastics in use.
It is currently produced in comparatively small quantities but is
widely used in aircraft interiors and as wire insulation. Poly-
imide combustion products are of concern primarily from the stand-
point of safety in an accidental combustion. Its formula (Figure
23) includes an aromatic group _R which can be varied to obtain
plastics of somewhat different properties. Synthesis and possible
groups of this polymer are described in the literature.
Sample
One polyimide film sample was studied. It was found by analysis
to be composed of 7.09 percent nitrogen, indicating that 71 milli-
grams of nitrogen per gram of sample should be accounted for in
its combustion products. The R group was not known for certain,
but CO "\0j— fits well with the nitrogen analysis.
Results
Thermal analysis. Differential thermal analysis was not performed.
Thermogravimetric analysis of polyimide in air (Figure 24) showed
a one-step degradation which did not begin until about 500 C. The
plastic was completely combusted by 650 C. This is consistant
with findings of Heacock & Berr,28 wh0 report, in addition, a one-
step degradation in helium beginning at the same temperature but
with only a 30-percent weight loss up to 1000 C.
Qualitative analysis. Infrared analysis of polyimide combustion
gas using a 10-cm path-length cell showed the presence of carbon
dioxide, carbon monoxide, ammonia, and cyanide (Figure 25). On
one occasion the scale-expanded infrared spectrum of polyimide
gaseous products showed evidence of nitrous oxide but no ammonia.
Gas chromatographic analysis of polyimide combustion gas showed
the fewest compounds of any plastic tested. On a six-foot-long,
1/4-in.-diameter Porapak Q column five peaks appeared: two due to
air, one due to carbon dioxide, one due to a combination of water and
cyanogen, and another due to hydrogen cyanide. All identifications
were made by mass spectrometry using an rtti MS-10. Nitric oxiae,
nitrogen dioxide, nitrous oxide, and ammonia were suspected but
could not be detected using a hydrogen flame gas chromatograph
detector. On a six-foot-long, 1/4-in.-diameter column of 10 per-
cent Carbowax 20 M on Chromosorb P several other gaseous products
appeared, benzene and benzonitrile being the only compounds identi-
fied by mass spectrometry. Polyimide produces a very small amount
of condensed liquid residue on combustion. This residue was ana-
57
-------
0 0
c c
0 0
Figure 23. Polyimide
58
-------
Figure 2k. Thermogravimetric analysis record of polyimide heated at 10 C/min in air
-------
FREQUENCY (CM1)
:0010000 5000 4000 3000 2500 2000 1800 1600 U00 1200 1100 1000 950 900 850 800 750
700 650
! 1
'CO
5X-
3*
fA
4=
=£10-
NM,
7 8 9
WAVELENGTH (MICRONS)
10 11
Figure 25. Infrared spectrum of polyimide combustion gas (10-cm path-length gas cell)
-------
lyzed using the GC-MS 30 combination and was found to be rich in
nitrogen-containing products (Table 13).
TABLE 13
IDENTIFICATION OF POLYIMIDE RESIDUE COMBUSTION PRODUCTS
GC Peak
Number
Molecular
Wei ght
Identi fi cation
1
Possibly water
2
87
N,N-Dimethyl acetamide
3
73
N-Methyl acetamide
101
Di acetami de
4
59
Acetami de
93
Ani1i ne
5
67 or 69
Unidentified
6
94
Phenol
7
Unidentified
8
129
Di acetyl ethyl ami ne?
9
—
Unidentified
10
Unidentified
Quantitative analysis. Carbon dioxide and carbon monoxide, quanti-
tated by infrared spectrophotometry, accounted for a very large
percentage (75 to 81) of the polyimide combusted, probably because
of the high temperature at which degradation begins to occur
(Table 14). Water vapor, not quantitated, would account for a
significant portion of the remaining plastic. The quantities in
Table 14 were obtained using the 0-to 800-C temperature range.
TABLE 14
CARBON DIOXIDE AND CARBON MONOXIDE FROM POLYIMIDE COMBUSTION
Air flow, cc/min
31
37
69
69
Oxygen flow, cc/min
26
0
0
0
Heating rate, C/min
5
5
5
50
Carbon dioxide*
1682
2037
1788
2005
Carbon monoxide
409
355
565
342
%Plastic accounted for 63.4 75.9 81.1 69.4
*The quantity of each combustion product is reported in
milligrams per gram of sample.
61
-------
Oxides of nitrogen (nitric oxide and nitrogen dioxide) were deter-
mined by the Saltzmun spectrophotometric method and were found to
be present in very low concentrations using the 0-to 1000-C temp-
erature range. Higher temperatures would probably be required to
produce significant quantities of nitrogen oxides. Note the
trend of nitrogen dioxide to increase with increasing air
supply (Table 15). Water and cyanogen elute at the same time on
the Porapak Q chromatograph column and other interferences are
possible, so this column was not used for quantitative work. Hydro-
gen cyanide was initially quantitated using a silver nitrate titra-
tion in the presence of potassium iodide, leading to an averaqe
value of 15 mg cyanide /g. Interference with this titration could
not be ruled out and the analyses were repeated using an Orion
Cyanide Ion Electrode. With it, 15 to 30 milligrams of cyanide
were determined per gram of sample (Table 15). Semiquantitative
analysis of benzonitrile indicated about 5 milligrams per gram
of sample. Compounds in the liquid residue were not quantitated.
TABLE 15
CYANIDE AND NITROGEN OXIDES FROM POLYIMIDE COMBUSTION
Air flow, cc/min 1
00
100
475
100
Oxygen flow, cc/min
0
20
0
0
Heating rate, C/min
5
5
5
50
Cyanide ion (as HCN)*
24.
3
29.
.7
21.
.6
18.9
Nitric oxide (NO)
0.
22
0.
,16
0.
.48
N.A.t
Nitrogen dioxide (NO2)
0.
20
0.
.53
0.
.71
N.A.t
*The quantity of each combustion product is reported in
milligrams per gram of sample.
+Not analyzed.
Discussion
Total accounting of polyimide nitrogen was not accomplished, al-
though significant quantities of cyanide and organic nitrogen com-
pounds were found. Nitrous oxide was not quantitated due to tech-
nical problems, but it is the oxide of nitrogen most likely to be
present under our combustion conditions. Experience with other
plastics, particularly Lopac^ and Barex*, has shown that molecular
nitrogen may well be a combustion product of nitrogen-containing
polymers, although it would not be detected in the presence of air.
As explained in the section on polyurethanes, the combination of
carbon monoxide and hydrogen cyanide is particularly hazardous in
accidental combustion.
62
-------
T(°C)
Figure 27* Thermogravime-ric analysis record of Lopac^ hea-ei at. 10 c/min in air
-------
FREQUENCY (CM')
2000010000 5000 4000 3000 2500 2000 1800 1600 1400 1200 1100 1000 950 900 850 800 750 700 650
Figure 28. Infrared spectrum of LopacR combustion gas (10-cm path-length gas cell)
-------
conditions. The endothermic nature of initial LopacR degradation
might limit Lopac'sR value in incinerators attempting to recover
thermal energy. However, further degradation of the endothermic
depolymerization products may be exothermic.
TABLE 16
CYANIDE FROM LOPACR COMBUSTION
Air Flow
(cc/min)
Oxygen Flow
(cc/min)
Heating Rate
(C/min)
Cyanide Ion
Concentration (mg/g)
100
0
5
6.64
100
20
5
7.54
475
0
50
3.51
67
-------
RARtxR
In Iroduc Li on
Barex^ is another plastic currently being test-marketed for soft-
drink bottles. It is manufactured by Vistron Corporation and is
reported to be an acrylonitrile-ethyl acetate-butadiene terpolymer.
The potential market for plastic soft-drink bottles in general is
in the billion bottle-per-year ranged and, since such bottles
would be used once and disposed, their use could have a significant
impact on the composition of solid waste.
Sample
The sample studied was Barex^-210, lot number 758370. It is com-
posed of 17.43 percent nitrogen. Since this plastic is still
being developed and tested, significant lot variations would not
be surprising.
ResuHs
Thermal analysis. Differential thermal analysis was not performed.
Thermogravimetric analysis of Barex^ in air (Figure 29) shows a
multi-step decomposition with only a 3-percent weight loss by
310 C. A rapid 24-percent weight loss occurs between 310 and 327
C and there is a more gradual 26-percent loss between 327 and
450 C. Another 11-percent weight loss occurs very gradually until
600 C, leaving a 36-percent residue at that temperature.
Qualitative analysis. Infrared analysis of Barex^ combustion gas
(Figure 30) in a 10-cm path-length cell shows bands due to C-H
groups, carbon dioxide, carbon monoxide, methane, ammonia, and
cyanide. Other combustion gases were not identified. Barex^
also produces a significant amount of liquid residue which chroma-
tographically shows about fifteen components on a five-foot-long,
1/4-in.-diameter column of 3 percent SE 30 (Figure 31). Tentative
identifications of these chromatogram peaks are listed in Table 17.
Of these compounds, benzonitrile is certainly present and a methyl
pyridine and a dimethyl pyridine are strongly indicated, although
isomers cannot be ruled out. Deficiencies in the mass spectra
obtained make identifications of the other compounds tentative.
Quantitative analysis. Ammonia was the only combustion product
quantitated. Using the 100-cc/m.in air flow and 5-C/min heating
rate 26 milligrams of ammonia were produced per gram of sample,
while using the same conditions but with 20-cc/min oxygen added,
15 milligrams ammonia were produced per gram of sample.
Discussion
Safe incineration is a major requirement for any plastic beverage
68
-------
100 200 300 400 500 600
TCC)
Figure 29- Therrr.ogr^v'imetric analysis record of Barex" neaped at 10 C/tnin in air
-------
FREQUENCY (CM')
2000010000 5000 4000 3000 2500 2000 1800 1600 1400 1200 1100 1000 950 900 850 800 750 700 650
WAVELENGTH (MICRONS)
Figure 30. Infrared spectrum of Barex-^ combustion gas (10-cra path-length gas cej—)
-------
-------
Dottle. The manufacturer of BarexR thus sponsored a study by the
Midwest Research Insitute, Kansas City, Missouri'30 in which 0.5
to 8 percent BarexR-210 was added to normal municipal refuse and
incinerator stack gases were monitored. Major stack gases were
carbon dioxide, water, and nitrogen, with no detectable levels of
acrylonitrile, methyl acrylate, acrolein, acrylic acid, or hydrogen
cyanide. Properly operated municipal incinerators should com-
pletely combust BarexR-210 releasing no undesirable gaseous pro-
ducts. However, the farther removed one gets from proper inciner-
ation conditions, the more likely is the generation of hydrogen
cyanide and other unidentified products. This is illustrated
both by Midwest Research Insitutue's finding of just detectable
(0.05 ppm) hydrogen cyanide, slightly increased nitrogen oxides,
and methyl acrylate and acrylonitrile on using an apartment-size
incinerator and by our own finding of large amounts of products
under very unfavorable combustion conditions.
TABLE 17
TENTATIVE IDENTIFICATION OF BAREXR RESIDUE CHR0MAT0GRAM PEAKS
Peak
Number
Molecular
Weight
Identification
1
48
Methyl mercaptan?
2
54
Butadiene
3
78
Benzene
4,5,6
74
Propionic acid
56
Acrolei n
7
68
Methyl butadiene*
55
Ethyl cyanide
8
69
Unidentified
9,10
93
Methyl pyridine*
11
107
Dimethyl pyridine*
94
Phenol
12,13
103
Benzonitrile
14,15
107
Dimethylpyridine*
16
109
Methoxypyri di ne*
17
108
Cresol
18
68 or 69
Unidentified
19
117
Methyl cyanobenzene
*Isomer unknown.
72
-------
UREA FORMALDEHYDE
Introduction
Urea formaldehyde is a thermoplastic resin used for electronic
components.
Sample
The sample was a wood-flour-fi1 led urea formaldehyde supplied by
Professor R. W. Heimburg and also used in his study of the
"Incineration of Plastics Found in Municipal Wastes". It was com-
posed of 23.53 percent nitrogen.
Results
Thermal analysis. Differential thermal analysis was not performed.
Thermogravimetric analysis in air (Figure 32) showed a three-step
degradation. Although a 3-percent weight loss occurred before
175 C, the first step was a gradual 11-percent loss between 175
and 260 C. This was followed by a steep 57-percent loss between
270 and 345 C. Above 345 C weight was gradually lost until the
plastic was completely combusted at 610 C.
Qualitative analysis. Infrared analysis of urea formaldehyde com-
bustion gas in a 10-cm path-length cell (Figure 33) showed the
presence of carbon dioxide, carbon monoxide, methane, ammonia, and
cyanide. This sample was run under quite inefficient combustion
conditions, producing large quantities of ammonia and cyanide.
Other gaseous products were not identified.
Quantitative analysis. Only cyanide ion was quantitated. Using
a 100-cc/min air flow and 5-C/min heating rate, 18.9 milligrams
cyanide ion were produced per gram of sample.
Discussion
This sample was sent to us by Professor Heimburg because it was
particularly toxic to plants and animals when burned using ineffi-
cient combustion conditions. His report states': "However, if
there was not enough premixing or a high enough afterburner temp-
erature, urea formaldehyde became the most hazardous and corrosive
material encountered." His study did not run other nitrogen-
containing polymers such as polyurethane or BarexR under similarly
deficient conditions, however. Cyanide and carbon monoxide are
the only acutely toxic compounds identified in either of our
studies. The fact that animals removed from exposure before
death recovered rapidly is evidence for this particular combination
of toxicants being responsible for the observed effects.
73
-------
T (°C)
Figure 32. Thermogravimetric analysis record of urea formaldehyde headed at 10 C/min in air
-------
FREQUENCY (CM')
2000010000 5000 4000 3000 2500 2000 1800 1600 1400 1200 1100 1000 950 900 850 ¦ 800 750 700 650
Figure 33* Infrared spectrum of urea formaldehyde combustion gas (10-cm path-length gas cell)
-------
PHENOL FORMALDEHYDE
Introduction
Phenol formaldehyde is a thermoset plastic and is among the older
resins now in use. It is used extensively in electrical equipment
and many studies of its combustion products have been made, mostly
in vacuum. According to Madorsky^, who quantitated approximately
fifteen pyrolysis products at four constant temperatures, the
resin is highly cross-linked because formaldehyde can react ortho,
meta, or para to the hydroxyl group of phenol.
Sample
The sample tested was a wood-flour-fi1 led Bakelite^ provided by
Professor R. W. Heimburg and also used in his study of "Incinera-
tion of Plastics Found in Municipal Wastes "J
Results
Thermal analysis. Differential thermal analysis was not performed.
Tfierniogravimetric analysis in air (Figure 34) showed a multi-step
decomposition. There was a gradual 3.5-percent weight loss by
175 C followed by a 20.5-percent loss between 175 and 360 C. There
was then a steady weight loss up to 620 C, at which temperature
30.5 percent of the plastic remained uncombusted. Burns & Orrell
have found four temperature ranges up to 400 C, where changes,
including curing and crosslinking, occur.31 Shebozake et al_ have
also observed that sample weight may affect the TGA curves of
phenol formaldehyde in air.32
Qualitative Analysis. Major combustion products of phenol formal-
dehyde determined by infrared analysis include carbon dioxide,
carbon monoxide, methane, and possibly anmonia (Figure 35). Cya-
nide concentration was below the detection limit of infrared
spectrophotometry. Minor combustion products were not identified.
Quantitative analysis. One analysis of phenol formaldehyde pro-
ducts for ammonia was made. Using the 100-cc/min air flow and 5-
C/min heating rate, 7,06 milligrams of ammonia were observed per
gram of sample. Cyanide ion was also quantitated as indicated in
Table 18. The last four values were obtained on a single heating
run. Cyanide ion is formed primarily during the early stages of
combustion, at the lower temperatures.
Discussion
Professor Heimburg found this plastic produced no ill effects in
76
-------
100
—J
—J
75-
o
LlI
* 50
I-
£
v°
25
100
200
300
400
500
600
T (°C)
Figure Tnermogrwimetric analysis record of phenol formaldehyde heated at 10 c/min in air
-------
FREQUENCY (CM')
2000010000 5000 4000 3000 2500 2000 1800 1600 U00 1200 1100 1000 950 900 850 800 750 700 650
WAVELENGTH (MICRONS)
Figure
Infrared spectrum of phenol formaldehyde combustion gas (10-cm pa~h-length gas cell)
-------
TABLE 18
CYANIDE FROM PHENOL FORMALDEHYDE COMBUSTION
Air Flow Oxygen Flow Heating Rate Temperature Cyanide Ion
(cc/min) (cc/min) (C/min) Range (C) Concentra-
tion (mg/g)
100
0
5
0-1000
3.51
100
40
5
0-1000
1.95
100
0
50
0-1000
2.09
100
0
5
0-550
2.09
100
0
5
550-850
1.14
100
0
5
850-1000
0.41
100
0
5
at 1000
0.15
either plants or animals when efficiently combusted. The parti-
cular phencl formaldehyde tested can form small amounts of nitrogen
compounds. No conclusion as to toxicity can be drawn pending
further analytical work.
79
-------
POLYETHYLENE
Iritroducti on
More polyethylene 1s produced than any other plastic, with the
high- and low-density materials totaling over six billion pounds
per year.4 For this reason and the fact
for packaging and other disposable goods
to solid waste is the largest of all the
consists of 85.7 percent
structure is:
that it is
its volume
piasti cs.
often used
contribution
Polyethylene
carbon and 14.3 percent hydrogen. Its
H
i
H
1
i
1
L
1
' L~
1
1
H
1
H
The high-density material differs from the low-density material in
having less branching of the chains. Combusting polyethylene to
carbon dioxide and water is not difficult, but incomplete combus-
tion leads to many hydrocarbon products. Madorsky has studied the
vacuum pyrolysis of polyethylene.5
Samples
Three polyethylene samples were studied: a low-density polyeth-
ylene powder, a high-density polyethylene powder, and a high-den-
sity polyethylene in pellet form.
Results
Thermal analysis. Differential thermal analysis was not performed.
Thermogravimetric analysis in air of each of the three polyethylene
samples (Figures 36 to 38) showed no weight loss until about 325 C
with each plastic completely decomposed by 500 to 550 C in a sin-
gle degradation step. Weight loss of the pellet sample was not
as smooth as that of the powder samples initially, perhaps because
of temperature gradients within the pellets.
Qualitative analysis. Infrared analysis of polyethylene combustion
gas in a 10-cm path-length gas cell showed carbon dioxide, carbon
monoxide, and methane to be present along with bands due to C-H,
alcohol, and carbonyl groups. , Separation of minor products on a
six-foot-ling, 1/4-in.-diameter column of Durapak n-Octane/Porasi1
C and identification by mass spectrometry revealed straight-chain
saturated and unsaturated hydrocarbons through Cg. Ethylene,
ethane, methanol, and acetaldehyde were separated on a column of
Porapak Q of the same dimensions as the Durapak column. Analysis
of the liquid residue left after polyethylene combustion by GC-MS
80
-------
T(°C)
Figure 36. Thermogravimetric analysis record of low-density polyethylene
(powder) heated at 10 C/min in air
-------
T(°C}
Figure 57 • Thermogravimetric analysis record of high-density polyethylene
(powder) heated at 10 C/min in air
-------
T(°C)
Figure 38- Thermogravimetric analysis record of high-density polyethylene
(pellets) heated at 10 c/min in air
-------
TABLE 19
COMBUSTION PRODUCTS OF LOW-DENSITY POLYETHYLENE (POWDER)
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Carbon dioxide*
88.
1610.
178.
Carbon monoxide
312.
171 .
110.
Methane
10.4
6.82
16.8
Ethylene
39.6
33.4
70.4
Ethane
4.65
2.83
11.2
Propylene
29.4
14.0
33.1
Propane
5.06
2.58
7.25
1-Butene
16.7
7.65
19.0
Butane
4.15
2.35
6.45
trans-2-Butene
5.68
4.23
9.0
cis-2-Butene
0.95
0.50
1.41
1-Pentene
9.13
4.30
11.8
Pentane
2.28
1 .06
3.35
1,3-Pentadi ene?
23.0
8.44
32.0
1-Hexene
10.0
4.92
14.7
2-Hexene
4.11
2.04
5.72
%P1asti c accounted for
32.3
60.7
33.8
*The quantity of each combustion product is reported in
milligrams per gram of sample.
using a five-foot-long, 1/4-in.-diameter column of low K' Durapak
resulted in a chromatogram composed of about ten peaks spaced in
a pattern characteristic of a homologous series of compounds.
Chromatographic retention times of the peaks matched well with
those of Cs-to C24-saturated hydrocarbons, but identifications
could not be made by mass spectrometry because the molecular ions,
and thus the molecular weights of the compounds, could not be
determined. Molecular ions of hydrocarbons often do not appear in
mass spectra, as they are very much less likely to form than frag-
mentation ions. A homologous series of unsaturated hydrocarbons
would actually seem more likely to form than one of saturated
hydrocarbons from polyethylene combustion.
Quantitative analysis. Carbon dioxide and carbon monoxide were
quantitated by infrared spectrophotometry. Most of the hydrocarbon
products were quantitated by gas chromatography on the same columns
84
-------
TABLE 20
COMBUSTION PRODUCTS OF HIGH-DENSITY POLYETHYLENE (POWDER)
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Carbon dioxide*
129.
753.
203.
Carbon monoxide
261.
179.
123.
Methane
15.2
7.75
17.7
Ethylene
32.6
31.0
49.6
Ethane
3.55
1.69
10.8
Propylene
37.7
20.4
37.2
Propane
4.92
1.92
8.5
1-Butene
18.7
9.79
18.5
Butane
3.21
1.44
4.51
trans-2-Butene
7.27
2.70
10.8
cis-2-Butene
1.08
0.54
1 .58
1-Pentene
10.8
4.51
13.1
Pentane
2.04
0.72
2.85
1,3-Pentadiene?
25.3
9.25
38.2
1-Hexene
11.2
3.96
16.5
2-Hexene
3.82
1.44
5.52
%Plastic accounted for
32.4
37.9
34.3
*The quantity of each combustion product is reported in
milligrams per gram of sample.
used for their separation, Methanol and acetaldehyde were not
quantitated. Quantitative results on seventeen combustion pro-
ducts of each of the three polyethylene samples are listed in
Tables 19 through 21. Thirty to sixty-five percent of the plastic
combusted can be accounted for by the identified products, with
the amount accounted for depending strongly on the amount of car-
bon dioxide produced. Unsaturated hydrocarbon products predomi-
nate over the saturated products. Interferences by other compounds
with the pentadiene peak have not been ruled out. Polyethylene
completely decomposed at such low temperatures that quantities of
its combustion products varied considerably. Sufficient data for
statistical analysis has not yet been obtained.
Discussion
Carbon monoxide was the only acutely toxic combustion product of
85
-------
TABLE 21
COMBUSTION PRODUCTS OF HIGH-DENSITY POLYETHYLENE (PELLETS)
AT SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/miri
5
5
50
Carbon dioxide*
213.
1842.
388.
Carbon monoxide
255.
173.
208.
Methane
13.1
6.09
17.6
Ethylene
49.0
30.5
52.9
Ethane
4.76
1.98
9.86
Propylene
34.8
11.8
35.3
Propane
5.11
1.78
7.69
1-Butene
18.1
5.77
16.0
Butane
3.63
1.33
4.08
trans-2-Butene
7.05
3.90
9.20
cis-2-Butene
1.01
0.43
1.26
1-Pentene
11.1
3.49
11.7
Pentane
2.22
0.84
2.41
1,3-Pentadi ene?
28.6
6.32
30.5
1-Hexene
12.4
2.55
16.1
2-Hexene
4.76
1.20
5.15
%Plastic accounted for
36.3
65.4
41 .5
*The quantity of each combustion product is reported in
milligrams per gram of sample.
polyethylene identified. The fuel or crude chemical value of the
liquid residue, however, should be of primary interest in solid
waste management. Under pyrolytic conditions, up to 90-percent
yields of the'liquid residue can be obtained.
86
-------
POLYPROPYLENE
Introduction
Polypropylene ranks fourth in the production of plastics with over
a billion pounds produced in 1971.^ Much of the polypropylene
produced is used for disposable or short-term-use goods, thus
contributing to the total solid waste problem. Polypropylene is
composed of 85.7 percent carbon and 14.3 percent hydrogen. Its
formula is
" 13 H
f-
K
I
H
C-
H
On complete combustion polypropylene would be expected to produce
only carbon dioxi-de and water. However, primary burning, open
burning, and inefficient incineration could result in production
of a variety of hydrocarbons. Madorsky has studied the pyrolysis
products of polypropylene,5
Sample
The sample studied was an isotactic (all methyls on the same side
of the carbon chain) polypropylene powder purchased from Cellomer
Associates, Inc., Webster, New York.
Results
Thermal analysis. Differential thermal analysis was not performed.
Thermogravimetric analysis of polypropylene in air (Figure 39)
showed a one-step degradation with the plastic completely combusted
by 440 C. Polypropylene begins breaking down at a lower tempera-
ture than polyethylene because, according to MadorskyS, alternate
carbons in the chain are tertiary making all the C-C bonds weaker.
Qualitative analysis. Carbon dioxide and carbon monoxide were
positively identified while methane and propylene were tentatively
identified by infrared analysis of polypropylene combustion gas
using a 10-cm path-length gas cell. Additional bands due to
hydroxyl and carbonyl groups were observed, possibly from methanol,
acetone, and/or acetaldehyde. Separation of gaseous products on
a six-foot-long, 1/4-in.-diameter Durapak n-Octane/Porasi1 C col-
umn followed by mass spectrometry showed a complete spectrum of C-j-
through Cg-hydrocarbons. Much of the polypropylene combusted
formed high-boiling liquid residue which has not yet been analyzed.
87
-------
T(°C)
Figure 59.
Thermogravimetric analysis record of isotactic polypropylene heated at 10 c/min in air
-------
TABLE 22
COMBUSTION PRODUCTS OF XSOTACTIC POLYPROPYLENE AT
SEVERAL COMBUSTION CONDITIONS
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Carbon dioxide*
131.
1195.
284.
Carbon monoxide
214.
284.
208.
Methane
22.8
25.4
151.
Ethylene
8.47
5.68
26.2
Ethane
0.98
0.46
9.77
Propylene
81 .0
21.2
314.
Propane
2.00
0.54
7.60
1-Butene
1.37
0.47
4.97
Butane
0.16
0.15
0.67
trans-2-Butene
7.12
2.63
30.4
cis-2-Butene
0.50
0.39
1.15
1-Pentene
2.30
0.31
4.50
Pentane
6.68
0.97
31.6
1,3-Pentadiene?
27.6
5.6
86.0
1-Hexene
12.0
1.70
34.4
%Plastic accounted for
30.0
51.3
86,9
*The quantity of each combustion product is reported in
milligrams per gram of sample.
Quantitative analysis. Available quantitative data on polypropy-
lene are listed in Table 22. These data represent single combust-
tion runs so it is not possible to comnent on their reproducibility.
Exceptionally large quantities of unsaturated hydrocarbons were
produced, particularly on fast burning. This may be due to the
low temperature at which polypropylene decomposes. Madorsky found
many of the same compounds on vacuum pyrolysis but in somewhat
different relative amounts.5
Discussion
Liquid residue from polypropylene combustion has some fuel value
which should be utilized. Combustion products thus far identified
are of low toxicity, except for carbon monoxide.
89
-------
POLYSTYRENE
Introduction
Polystyrene is used a great deal for packaging and disposable goods
and contributes significantly to the plastics found 1n solid waste.
It is composed of 92.3 percent carbon and 7.7 percent hydrogen,
and has the following formula: — —
Its products of combustion have been studied in vacuum by Madorsky5,
?nd it has been found that in vacuum styrene accounts for about
75 percent of the polymer degraded. With respect to incineration,
more emphasis has been placed on studying soot generation by poly-
styrene than on identifying gaseous products. Soot generation is
a fairly serious problem with aromatic polymers.
Sample
One polystyrene powder sample was obtained from Cellomer Associates,
Inc., Webster, New York.
Results
Thermal analysis. Differential thermal analysis was not performed.
TlWmogravimetric analysis of oolystyrene (Figure 40) showed a
one-step degradation beginning just before 300 C with the plastic
completely combusted at about 450 C.
Qualitative analysis. Separation of polystyrene combustiop pro-
ducts on a six-foot-long, 1/4-in.-diameter Porapak Q column showed
peaks due to methane, ethylene, ethane, propylene, propane,
1-butene, and possibly methanol and acetaldehyde. GC-MS analysis
of polystyrene combustion gas using a five-foot-long, 1/4-in.-dia-
meter 3 percent SE 30 column was sufficient to identify benzene,
toluene, ethylbenzene, styrene, and two isomers of methylstyrene.
Polystyrene liquid residue, a considerable amount of which formed
during combustion, was also analyzed by GC-MS using the SE 30
column. Twenty spectra were taken and about fifteen compounds
identified (Figure 41 and Table 23).
Quantitative analysis. Quantitative analysis was not done on
polystyrene comEustion products.
90
-------
T(°C)
Figure 1+0^-, Thermogravi:;.cx-j u-uuxysis recora ux polystyrene neatea at 10 C/min in air
-------
ro
TIME -
Figure Ul. Chromatogram of polystyrene liquid residue on a 3 percent SE 30 column
-------
Discussion
The major component of polystyrene combust/on in both the gas and
liquid phases is the monomer styrene, alti/jugh many other compounds
are present in significant amounts. Ther is considerable potent-
ial for monomer recovery, as mentioned in/a recent TRW report.33
Aromatic liquid-residue components are ch/.mically quite interesting,
as well as potentially toxic.
TABLE 23
IDENTIFICATION OF POLYSTYRENE LIQUID RESIDUE CHROMATOGRAM PEAKS
i
Peak
Number
Molecular
Weight
Identi fi cation
»
1
78
1
Benzene
2
92
Toluene
3
106
Ethylbenzene
4,5
104
Styrene
6
106
Benzaldehyde
118
Methylstyrene
7
94
Phenol
8
118
Methyl styrene
9
120
n-Propyl benzene
10
116
Indene
11
120
Acetophenone
12
130
Methyl indene
13
—
Unidentified
14
128
Naphthalene
15
134
Cinnamyl alcohol
16
142
Methyl naphthalene
17,18
154
Biphenyl or Acenaphthene
19
168
Methylbiphenyl
20
182
Diphenylethane
93
-------
POLYCARBONATE
Introduction
Polycarbonate is a specialty thermoplastic developed for its good
thermal stability, high strength, and clarity. Its applications
include electronic components and window material. Although not a
high-production plastic, its combustion products are of interest
because they contribute to the overall solid waste disposal pro-
blem and also to the toxicity of combustion products during acci-
dental fires in buildings or aircraft. The formula of bisphenol
A-polycarbonate, shown in Figure 42, consists of 75.58 percent
carbon, 5.55 percent hydrogen and 18.87 percent oxygen.
Samples
Three polycarbonate samples in commercial pellet form were studied.
Results
Thermal analysis. Differential thermal analysis records in helium
and in air of one polycarbonate sample are shown in Figure 43.
Altnough this plastic melts in a combustion furnace at about 370 C,
no melting endotherms are evident. Small endotherms at 475 C are
the first significant features of the DTA records, followed by
much stronger endotherms at 520 C. Many volatile and high-boilinq
products can be collected at temperatures corresponding to these
two endotherms. Above 520 C, the reactions are strongly exothermic
in air, due in part to formation of carbon monoxide and carbon dioxide,
and less strongly exothermic in helium, DTA records of the other
two polycarbonate samples were similar to the one shown, except for
weaker 520-C endotherms in air,
Thermogravimetric analysis records of another polycarbonate sample
are shown in Figure 44. In helium, a one-step degradation is ob-
served with a 75-percent weight loss between 420 and 520 C. The
plastic also begins to break down at 420 C in air, losing approx-
imately 55 percent of its weight by 480 C. Another 17 percent of
the weight is lost more gradually between 480 and 545 C, with the
remaining 28 percent being lost by 600 C. Thus, the 75-percent
weight loss occurring in one step in helium, occurs in two steps
in air, indicating "somewhat different reaction mechanisms.
Several studies of the mechanism of degradation of polycarbonate
have been published, Davis and Golden have shown that thermal de-
gradation in vacuo occurs by random chain scission^, and if vol-
atile products are continuously removed, the polymer is rapidly
cross linked to form a gel. 35 [_ee has proposed schemes for the
decomposition based on products identified at 475 C under vacuum
and in air, but pointed out that above 450 C a complicated mixture
94
-------
Figure k2.
CH3 / v
{-(7)-o
ch3 >—f
c
II
Bisphenol A-polycarbonate
-------
AT
100 200 300 400
T (°C)
500
600
Figure 1+3. Differential t/nermal analysis records of polycarbonate- heated at 10 C/min
in helium (broken curve) and air (smooth curve)
-------
T(°C )
Figure kb. Thermogravimetric analysis records of polycarbonate heated at 5 c/min
in helium (broken carve) and air (smooth curve)
-------
of reactions is rapidly taking place,36
Qua!itative analysis. Gaseous combustion products of polycarbon-
ate were separated on a six-foot long, 1/4-in.-diameter Porapak Q
column and on a 13-foot-long, 1/4-in.-diameter column of 5 percent
Squalane on Chromosorb P. The former column separated C] through
C4 hydrocarbons, methanol, and acetaldehyde. The latter column
separated benzene, toluene, ethylbenzene, and styrene. These vol-
atile products along with carbon monoxide and carbon dioxide,
which were identified by infrared spectroscopy, account for 40 to
60 percent of the polycarbonate combusted. The remainder of the
plastic is accounted for as liquid residue, some components of
which were identified using the GC-MS system with a 13-in.-long,
1/4-in.-diameter low K'Durapak column. A chromatogram of poly-
carbonate products on this column (Figure 45), shows about thirty
compounds, many of which are identified in Table 24. Peaks 18-28
were observed on the mass spectrometer total ion current monitor
as they were present in large enough quantities to produce off-
scale peaks. Although quantitative work was not done on the resi-
due, phenol and substituted Dhenols seem to predominate. Lee has
identified some of the higher-boiling compounds, including isopro-
pen^lphenol, isopropylphenol, diphenyl carbonate, and bisphenol
Quantitative analysis. During the study of polycarbonate a modi-
fication of the combustion furnace temperature controller changed
the temperature range from 0 to 800 C to 0 to 1000 C. Polycar-
bonate data using both temperature ranges are thus available,
althougn most of the study was done using the expanded range. In
all cases the temperature was programmed at 5-to 50-C/minute after
initial rapid heating from room temperature to 350 C. Table 25
lists 0-to 800-C data for one sample while Table 26 lists data
for the three polycarbonate samples using the 0-to 1000-C tempera-
ture range. Data at the 5-C/min and 50-C/min heating rates with
a 100-cc/min air supply can be compared for the two temperature
ranges. The narrower range results in approximately the same
amount of carbon dioxide, but much more carbon monoxide and slightly
less residue. About two-thirds of the hydrocarbon values are com-
parable for the two ranges, methane and benzene showing the largest
deviations.
Although there appear to tie trends within each temperature range
which can be explained on the basis of combustion efficiency,
these trends do not seem to hold between temperature ranges. For
example, at 5 C/min, results for the 0-to 800-C range indicate
that as the oxygen supply increases, quantities of carbon dioxide
and carbon monoxide increase and the amount of liquid residue
decreases. For the O-to 1000-C. range^ however, there is more
residue and more carbon monoxide at the higher oxygen supply but
less carbon dioxide, indicating less efficient combustion, as
98
-------
TIME —*
Figure . ^'r^omatogram of polycarbonate liquid residue
on a low K' Durapak column
-------
TABLE 24
IDENTIFICATION OF POLYCARBONATE RESIDUE CHROMATOGRAM PEAKS
Peak Number Molecular Weight Identification
1
78
Benzene
2
92
Toluene
3
106
Ethylbenzene or Xylene
4
104
Styrene
5
102
Phenyl acetylene
6
120
Isopropyl benzene or Ethyl
toluene
7
118
Isopropenyl benzene
8
118
Indan (2,3-Dihydroindene)
9
116
Indene
10
134
sec-Butyl benzene or isomer
11
132
Methyli ndan
12
132
Cinnamaldehyde or Vinyl
benzaldehyde
13
130
Methyli ndene
14
128
Naphthalene
15 .
136
O-Phenylene cyclic carbonate
146
Dimethylindan or Ethylindan
16
134
Methylacetophenone or Cymene
144
2-Naphthol ,3-Phenylfuran or
2-Vinyl benzofuran
17
142
Methyl naphthalene
18
94
Phenol
19
154
Biphenyl
168
Methylbiphenyl
170
Diphenyl ether
20
108
Cresol
152
O-Vanillin
21
122
Ethyl phenol
166,168
Unidentified
182
2-Ethylbiphenyl
184
Phenyl-p-tolyl ether
22
136
Isopropyl phenol
166
Unidentified
182
Unidentified
196, 198
Unidentified
23
150
Unidentified
210, 212
Unidentified
24
228
Bisphenol A
240
Unidentified
25
198?
Uni denti fied
242
Uni denti fied
26
256
Unidentified
27
270
Uni dentified
28
284
Unidentified
100
-------
TABLE 25
COMBUSTION PRODUCTS OF POLYCARBONATE AT SEVERAL COMBUSTION CONDITIONS
(800 C maximum)
Air flow, cc/min
100
100
100
200
Oxygen flow, cc/min
0
. 0
20
0
Heating rate, C/min
5
50
5
5
Carbon dioxide*
960
985
1010
1100
Carbon monoxide
610
270
780
1125
Methane
12.2
39.8
14.3
17.8
Ethylene
0.82
1.28
0.99
1 .07
Ethane
0.71
0.94
0.87
0.94
Propylene
0.31
0.50
0.39
0.33
Propane
0.67
0.15
0.15
0.15
Butane
0.20
0.19
0.91
0.11
Methanol
1.69
1.72
4.05
1 .71
Acetaldehyde
1.03
0.86
1.83
0.96
Benzene
3.49
1.48
2.05
2.44
Toluene
1.16
0.47
0.83
1.18
%Plastic Accounted for
63
51
72
89
%Residue (by difference)
37
49
28
11
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
TABLE 26
COMBUSTION PRODUCTS OF POLYCARBONATE AT SEVERAL COMBUSTION CONDITIONS
(1000 C maximum)
Air flow, cc/min
100
100
100
Oxygen flow, cc/min
0
40
0
Heating rate, C/min
5
5
50
Carbon dioxide*
1146
157
747
j-
36
991 ±
34
Carbon, monoxide
395
+
36'
425
±
34
76 ±
12
Methane
27.2
+
1.0
9.5
+
1.6
25.9 ±
0.2:
Ethylene
1.57
±
0.27
1.31
+
0.16
2.94 ±
1 .67
Ethane
0.92
+
0.08
0.80
+
0.02
0.99 ±
0.11
Propylene
0.82
+
0.33
0.56
a.
0.12
1.24 ±
0.81
Propane
0.23
+
0.11
0.14
±
0.04
0.25 ±
0.14
Methanol
1.48
+
0.18
1.89
±
0.32
1.07 ±
0.11
Acetaldehyde
0.72
+
0.38
0.90
+
0.41
0.58 ±
0.34
1-Butene
0.26
+
0.12
0.20
+
0.07
0.56 ±
0.33
Butane
0.08
+
0.07
0.05
+
0.05
0.09 ±
0.08
Benzene
1.97
+
0.08
2.48
u.
0.69
3.75 =
0.22
Toluene
1.10
+
0.12
1 .31
+
0.22
3.33 _
0.25
Ethyl benzene
0.67
+
0.18
0.69
±
0.03
1.90 =
0.15
Styrene
0.10
+
0.03
0.10
0.01
0.31 ±
0.02
%Plastic accounted for
60
49
43
%Residue (by difference)
40
51
57
*The quantity of each combustion product is reported in milligrams per gram of samcie
-------
explained in the Methodology section.
Quantities of products for the three polycarbonates tested using
the 0-to 1000-C range were so similar that results were averaged
(Table 26), Standard deviations were calculated which include in
this case differences between the three plastics as well as experi-
mental error and the randomness of the combustion process. One
sample contained some filler material which contributed excess
hydrocarbons, so the averages for some minor products are weighted
toward that sample. Where deviations are larger than 25 percent,
such as for propylene and butane, that sample is responsible.
Carbon monoxide, carbon dioxide, and methane are the major vola-
tile products. The latter two show significant decreases in the
presence of excess oxygen. The faster heating rate allows most
of the reaction with oxygen to occur at higher temperatures and
thus favors production of carbon dioxide over carbon monoxide.
The amounts of aliphatic hydrocarbons decrease with chain length.
The unsaturated aliphatic compound is consistently present in
greater amount than the corresponding saturated compound. Rela-
tively small amounts of volatile aromatic hydrocarbons were
detected. This may be due to the fact that the large amount of
liquid residue consists of phenol and substituted phenols in
which the lower-boiling aromatics are quite soluble. The percent
plastic accounted for was calculated as the sum of "the percents
of the volatile products, assuming that carbon dioxide and carbon
monoxide are formed from half of the 16 percent oxygen in poly-
carbonate and that the remainder of these compounds are formed by
reaction of the carbon skeleton of the plastic with oxygen in the
air. Volatile products, identified in quantities less than 2 milli-
grams per gram of sample but not otherwise quantitated, include
methyl acetylene, acetone, isobutane, cis-2-butene, pentane, cyclo-
pentene, and xylene.
Table 27 shows quantities of the combustion products of one poly-
carbonate sample obseryed within temperature ranges selected on
the basis of the TGA curve. Totals for this sample are quite
representative of the two unfilled polycarbonates. As expected
carbon dioxide and carbon monoxide form primarily in the last
stage, but also are significantly present in the first stage.
This is the first tested plastic which formed these compounds at
such low temperatures, indicating decomposition of the carbonate
linkages. If all the oxygen of carbon dioxide and carbon monoxide
in the first temperature range came from carbonate linkages, about
half the oxygen of the plastic would be accounted for. Only the
quantities of methane did not add up to a total expected from
Table 26. The percent plastic accounted for is about the same.
The lowest range is quite deficient in aliphatic hydrocarbons, but
otherwise no trends are evident. Aromatic hydrocarbons are formed
primarily at higher temperatures, that is, after the 520-C
endotherm.
103
-------
TABLE 27
POLYCARBONATE COMBUSTION PRODUCTS WITHIN SEVERAL TEMPERATURE RANGES
Compound
<475 C
475 -
500 C
500 -
550 C
550 -
1000 C
Total
Carbon dioxide*
90
60
133
997
1280
Carbon monoxide
10.3
14.6
60
248
333
Methane
2.25
2.48
5.69
3.75
14.2
Ethylene
0.09
0.31
0.39
0.33
1 .12
Ethane
0.057
0.19
0.36
0.12
0. 73
Propylene
0.U95
0.17
0.085
0.026
0.37
Propane
0.022
0.067
0.03
0.12
Methanol
0.093
0.43
0.36
0.14
1 .02
Acetaldehyde
0.092
0.10
0.06
0.085
0.34
1-Butene
0.062
0.038
0.008
0.042
0.15
Butane
0.001
0.004
o.ooi
0.004
0.01
Benzene
0.045
0.077
0.72
1.06
1 .90
Toluene
0.066
0.19
0.46
0.18
0.90
Ethylbenzene
0.013
0.088
0.21
0.17
0.48
Styrene
0.006
0.008
0.014
0.036
0.06
^Plastic accounted for 59
^Residue (by difference) 41
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
Discussion
Polycarbonate is typical of the polymers containing only carbon,
hydrogen, and oxygen in that, on inefficient combustion, it under-
goes random chain scissions to form a large amount of liquid. The
phenolic compounds are expected but naphthalenes and indenes had
not been previously reported and the mechanism of their formation
is not completely understood,
105
-------
POLYPHENYLENE OXIDE
Introduction
Polyphenylene oxide is a linear polymer, the formula of which is
shown in Figure 46. It consists of 79.3 percent carbon, 7.5 per-
cent hydrogen, and 13.2 percent oxygen. It is commercially provided
in the pure form and in a modified form, both being self-extinguish-
ing and non-dripping and having continuously usable temperature
ranges extending to above 100 C.21>37 Military, aircraft, and
household uses for this plastic have been developed and promise to
increase because of its good thermal, electrical, and mechanical
properties. Preliminary studies on this plastic were done under a
previous grant.20
Samples
Of the four commercial samples tested, sample 1 consists of poly-
phenylene oxide with a small amount of titanium dioxide whitener,
and samples 2 through 4 are modified polyphenylene oxides with
either carbon or glass fillers.
Results
Thermal analysis. Differential thermal analysis records of poly-
phenylene oxide, heated at 5 C/min in helium and in air atmospheres,
are shown in Figure 47. In helium two endothermic peaks appear,
one at 425 C and one at 460 C. In air the 460-C endothermic peak
is preceded and followed by exothermic peaks which may represent a
continuous exothermic reaction (oxidation of carbon) on which an
endothermic reaction (depolymerization) is superimposed. This is
consistent with the observation that the plastic begins to char
at about 375 C. A thermogravimetric analysis record of the same
plastic, heated at 5 C/minute in an air atmosphere, is shown in
Figure 48. The major weight lo,ss occurs in two steps. The first
begins at about 375 C and continues to 450 C, accompanied by a 56-
percent weight loss. This is followed by a transition and a 10-
percent weight loss from 450 to 510 C. The second major step begins
at about 510 C and accounts for combustion of the remaining 34 per-
cent of the plastic. The temperature at which the plastic is com-
pletely combusted depends on the air supply. Correlating DTA and
TGA results, the DTA endotherms must correspond to depolymerization,
rather than melting, as they occur at temperatures where large
weight losses have occurred. The 460-C endotherm corresponds to
the end of the first step in the TGA curve. The TGA transition is
not clearly defined in the DTA record but may be represented by a
shoulder on the large exothermic peak. This large exothermic DTA
peak corresponds to the second step of the TGA curve.
106
-------
C H:/
-1 n
Figure k6. Poly-2,6-dimethyl-l,i+-phenylejie oxide
107
-------
AT
Ex
100 200 300
Figure 47. Differential therma^. analysis
in helium (broken curve) and air (smooth
V.-A /
V
/
400 500 600
TCC)
ecords of polyphenylene oxide heated £~ ~ C/rnin
rvp ^
-------
T(°C)
Figure 48- Thermogravimetric analysis record of polyphenylene oxide heated at 5 C/min in air
-------
A similar situation exists for modified polyphenylene oxide as
shown in Figures 49 and 50. In both oxidizing and inert atmospheres
one significant endothermic peak occurs at 435 C and corresponds
to the first step shown on the TGA record, a depolymerization with
a 64-percent weight loss. Since the DTA records of modified poly-
phenylene oxide and polyphenylene oxide in helium were made using
the same sample weights, a comDarison indicates that depolymeriza-
tion of modified polyphenylene oxide is more stongly endothermic.
The shoulder on the large exothermic peak in Figure 49 could corres-
pond to the TGA transition, while the large exothermic DTA peak
above 500 C corresponds to the second step of the TGA curve.
The complexity of the combustion process makes interpretation of
all features of the DTA and TGA records difficult. Major emphasis
has been placed on interpreting those features of the DTA records
involving generation of products, with little emphasis on deter-
mining melting, softening, or glass transition temperatures which
may be found elsewhere.38
Qualitative analysis. Carbon dioxide, carbon monoxide, and methane
were identified oy infrared analysis of the combustion gas using
a 10-cm path-length gas cell. Strong bands characteristic of
aromatic compounds also appeared in the infrared spectrum. Further
identifications were made by mass spectrometry after separation of
individual products by gas chromatography.
Gaseous polyphenylene oxide combustion products separated on a
six-foot-long, 1/4-in.-diameter column of Porapak Q include hydro-
carbons through butane plus methanol and acetaldehyde (Figure 51).
Figure 52 shows a chromatogram of gaseous polyphenylene oxide com-
bustion products on a 12-foot-long, 1/4-in.-diameter column of 5
percent Squalane on Chromasorb P. A total of seventeen gaseous
products were eluted in the same order as that presented in Table
29 beginning with methane and continuing through styrene. As can
be seen from the attenuations indicated in Figure 52 aromatic com-
pounds are present in substantially greater quantities than
aliphatics. All four samples show the same qualitative picture
except that samples 2 through 4 have a few additional small peaks.
No attempt has been made to identify some of the smaller peaks in
the Squalane chromatogram.
Between 400 C and 475 C, water and dense yellow-brown fumes are
produced which condense in the combustion tube on leaving the
heating zone. Much of the weight loss in the first step results
from this condensate. Any of the remaining organic volatile mater-
ial boils at too high a temperature to be eluted in a reasonable
time chromatographically under column conditions used for the vola-
tile products. Other investigators found that, in vacuum, products
volatile at room temperature represent less than 4 percent of the
total.39
110
-------
AT
100 200 300
400
T(°C)
500
600
Figure b9- Differential thermal analysis records of modified polyphenylene
oxide heated at 5 C/min in helium (broken curve) and air (smooth curve)
111
-------
Figure 50. Thermogravimetric analysis record of modified polyphenylene oxide heated at 5 c/min in air
-------
Figure 51« Chromatogram of polyphenylene oxide combustion products on a Porapak Q column
-------
Figure 52. Chromatogram of polyphenylene oxide combustion products
on a column of 5 percent squalane on Chromosorb P
-------
Although the primary aim of this study was to analyze the volatile
combustion products, 1t was decided to identify some components of
the condensate because of their large amount, pungent odor, and
indications from preliminary work that they could be toxicologi-
cally more significant than the identified volatile products.
Infrared analysis of the condensate, after removal of water, indi-
cates that it is a complex mixture of aromatic compounds. It
includes small amounts of toluene, ethylbenzene, and styrene so
that cited quantitative figures for these compounds must be accepted
as minima.
The GC-MS 30 instrumentation with a 13-in.-long, 1/4-in.-diameter
column of low K' Durapak was used to identify components of the
liquid residue. About thirty peaks appeared in the chromatogram
(Figure 53). The numbered peaks are characterized or identified
in Table 28. Many peaks represent compounds present in small
amounts. There are only four major peaks and these indicate the
presence of 2, 6-xylenol, 0-cresol , 2,4-xylenol and trimethylphenol.
The isomers were determined by infrared spectroscopy, as mass
spectra of isomers are nearly identical. Although no quantitative
work was done on the liquid residue components, the quantities of
the four major compounds are estimated to be ten to one hundred
times as great as those of the other components of the condensate.
Quantitative analysis. Table 29 shows quantities of identified
gaseous products from each of the four samples run using a tempera-
ture programming rate of 5 C/min from 350 to 800 C after initial
rapid heating to 350 C. There was sufficient air, when integrated
over the entire run, to convert all the carbon to carbon dioxide,
as explained in the Methodology section. Actually there was
a slightly deficient oxygen condition. All results are given in
milligrams per gram of plastic. Carbon dioxide and carbon monoxide,
as expected, are the major volatile combustion products. Although
the ratio of the two compounds varies considerably, significant
amounts of carbon monoxide are produced by all the samples.
Straight-chain saturated and unsaturated hydrocarbons, except
methane, are present in trace amounts, with an unsaturated compound
occurring at about twice the concentration of the corresponding
saturate. The three substituted aromatic compounds rank with
methane as the principal volatile hydrocarbon constituents of the
combustion products. Residue represents only inert material remain-
ing in the combustion boat after a run and does not include vola-
tile products which condense in the end of the combustion tube or
in the collection bag. It can be seen that Sample 1, the unmodified
polyphenylene oxide, differs slightly from the other samples, which
are modified polyphenylene oxides. Samples 2 and 3 show nearly
identical products and differ from Sample 4, which forms a large
amount of inorganic residue. Volatile products listed in Table 29
account for about 50 percent of the weight of each polymer combusted.
Water (vapor and liquid), which was not quantitated, and high-
115
-------
TIME
Figure 53* Chromatogram of polyphenylene oxide liquid residue on a low K' Durapak column
-------
boiling organic compounds account for the remainder of the plastic.
TABLE 28
IDENTIFICATION OF POLYPHENYLENE OXIDE RESIDUE CHROMATOGRAM PEAKS
Peak
Molecular
Identification
Number
Weight
1
78
Benzene
2
92
Toluene
3
106
Ethylbenzene or Xylene
4
104
Styrene
5
120
Tri methyl benzene*
6
118
Indan
7
116
Indene
8
Unidentified
9
Unidentified
10
132
Phenyl propynyl ether*
11
134
o-Cymene*
128
Naphthalene
12
136
o-Cresyl ethyl ether*
13
146
Dimethylindan
14
150
Unidentified
15
—
Unidenti fied
16
142
Methylnaphthalene
17
122
2,6-Dimethylphenol
18
108
o-Cresol
19
136
Trimethylphenol
20
122
2,4-Dimethyl phenol
166
2-(x-Xylyloxy) ethanol?
21
136
Trimethylphenol
22
196
Unidentified
*Or isomer.
Sample 1 was employed to test the effect of combustion conditions
on amounts of products generated. Several runs were made using
the 5-C/min heating rate and varying amounts of air. The first
air supply was about half that required to convert all the carbon
of the plastic to carbon dioxide. The second air supply, when
integrated over the entire run, was just sufficient to convert all
the carbon to carbon dioxide. The addition of pure oxygen to the
third air supply provided an excess. Results are shown in Table
30. The run with deficient oxygen shows the least amount of car-
bon monoxide. This is explained by the fact that the plastic
requires about a one-half-hour exposure at 800 C, the maximum
temperature of our apparatus, to combust completely in this air
117
-------
TABLE 29
QUANTITIES OF COMBUSTION PRODUCTS FROM SEVERAL POLYPHENYLENE OXIDE PLASTICS
Product
Sample
1
Sample
2
Sample
3
Sample
4
Carbon dioxide*
838.
503.
572.
337.
Carbon monoxide
457.
425.
377.
330.
Methane
18.2
9.09
9.63
5.33
Ethylene
0.58
1.27
1.01
0.54
Ethane
0.80
0.65
0.64
0.39
Propylene
0.30
0.50
0.40
0.17
Propane
0.22
0.26
0.24
o.n
Methanol
1.52
1 .60
1.46
0.97
Acetaldehyde
1 .13
1.77
1.34
0.98
1-Butene
0.29
0.66
0.57
J. 20
Butane
0.17
0.24
0.23
0.08
1-Pentene
0.29
0.33
0.27
0.07
Pentane and pentadiene
0.32
0.31
0.27
0.08
Hexane
0.09
0.10
0.09
0.02
1-Hexene
0.22
0.29
0.23
0.04
Benzene
1 .76
3.38
3.17
2.14
Toluene
22.9
34.6
34.8
27.8
Ethyl benzene
12.4
22.9
25.3
16.2
Styrene
21 .5
72.0
64.8
57.4
Resi due
14.0
3.0
1.0
260.0
%Plastic accounted for
52
47
46
61
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
VARIATION OF
TABLE 30
POLYPHENYLENE OXIDE COMBUSTION
PRODUCTS WITH
AIR SUPPLY
Product
Deficient
°2
Suffi ci ent
02
Excess
°2
Carbon dioxide*
887.0
838.0
830.0
Carbon monoxide
325.0
457.0
452.0
Metnane
20.8
18.2
16.5
Ethylene
0.99
0.58
0.80
Ethane
0.82
0.80
0.68
Propylene
0.40
0.30
0.35
Propane
0.23
0.22
0.19
Methanol
1.68
1 .-52
2.65
Acetaldehyde
0.97
1 13
1.87
1-Butene
U. 65
U.
0.31,
Butane
0.15
0.17
w • 1 O
1-Pentene
0.27
0.29
0.22
Pentane and pentadiene
0.25
0.32
0.19
1-Hexene
0.29
0.22
0.19
Hexane
0.09
0.09
0.06
Benzene
2.76
1 .76
2.29
Toluene
23.3
22.9
21.3
Ethyl benzene
10.4
12.4
10.7
Styrene
18.3
21 ..5
15.3
*T(ie quantity of each combustion product is reported in milligrams per gram of sample.
-------
supply. Since the amount of oxygen at any given time is less, the
time required for combustion is much longer. Using the other two
air supplies, the plastic is combusted at or before 800 C. It is
significant that the amount of carbon monoxide generated seems to
remain constant above a certain air supply. Minor fluctuations are
observed 1n the other products, but they are so small that it can-
not be determined whether they represent real trends, experimental
error, or the normal deviations to be expected in the combustion
process.
Table 31 shows the influence of heating rate on amounts of products
generated. Using the sufficient airflow in each of the runs, the
plastic was heated at the indicated temperature rate to 800 C. The
study using the faster heating rate involved longer 800-C exposure
before combustion was complete and favored the production of carbon
dioxide rather than carbon monoxide. It is characteristic of the
unsaturated-saturated aliphatic pairs that the unsaturated compound
increases two to threefold at the higher heating rate, while the
corresponding saturated compound remains about the same. Again,
other trends are of a very minor nature.
Table 32 shows the quantities of the products produced in four temp-
erature ranges selected on the basis of the TGA curve. The 5-C/min-
ute heating rate and sufficient air supply were used. The first
two temperature ranges cover the first step of the breakdown process,
the third range covers the transition, and the last range covers
the second step. As expected, carbon dioxide and carbon monoxide
are mainly produced at higher temperatures, and their amounts
account for nearly all the weight loss in the second step of the
TGA curve. Aliphatic hydrocarbons except methane peak during the
third temperature range. Aromatic hydrocarbons show no real trend
as benzene peaks in the last range, toluene in the second, and
ethyl benzene is about equally divided among the last three ranges.
The amounts of volatile products from the first three temperature
ranges should account for 0.6 gram of plastic, as indicated by the
60 percent weight loss on the TGA record. Considerably less of the
plastic is actually accounted for by these products due to the
generation of liquid residue.
Discussi on
It is evident that, of the volatile combustion products identified,
carbon monoxide presents the only significant health hazard. The
substituted phenols identified in the condensate are all irritants
and have significant vapor pressures at room temperature. Any
contribution of the condensate to the overall toxicity of the com-
bustion products could best be evaluated from animal exposure data.
120
-------
TABLE 31
EFFECT OF HEATING RATE ON COMBUSTION PRODUCTS OF POLYPHENYLENE OXIDE
Product
5 C/min
50 C/min
Carbon dioxide*
838.0
908.0
Carbon monoxide
457.0
179.0
Methane
18.2
22.3
Ethylene
0.58
1.77
Ethane
0.80
0.76
Propylene
0.30
0.86
Propane
0.22
0.26
Methanol
1.52
0.94
Acetaldehyde
1.13
0.97
1-Butene
0.29
0.67
Butane
0.17
0.22
1-Pentene
0.29
0.51
Pentane and pentadiene
0.32
0.44
1-Hexene
0.22
0.64
Hexane
0.09
0.10
Benzene
1.76
3.03
Toluene
22.9
17.6
Ethyl benzene
12.4
7.07
Styrene
21 .5
21.4
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
TABLE 32
VARIATION OF POLYPHENYLENE OXIDE COMBUSTION PRODUCTS WITH TEMPERATURE
Product
350 -
400 C
400 -
450 C
450 -
510 C
510 -
800 C
Carbon dioxide*
3.5
23.8
89.7
780.0
Carbon monoxide
8.5
23.3
65.5
340.0
Methane
0.057
1.16
5.85
11.7
Ethylene
0.017
0.11
0.31
0.27
Ethane
0.041
0.14
0.46
0.18
Propylene
0.008
0.10
0.24
0.04
Propane
0.008
0.08
0.17
0.01
Metnanol
0.015
0.51
1.23
0.31
Acetaldenyde
0.54
0.30
0.35
0.01
1-Butene
0.053
0.11
0.15
0.03
Butane
0.016
0.09
0.11
0.01
1-Pentene
0.008
0.09
0.17
0.02
Pentane and pentadiene
0.013
0.09
0.15
1-Hexene
0.008
0.07
0.17
0.03
Hexane
0.005
0.03
0.06
Benzene
0.087
0.37
0.43
0.79
Toluene
1 .02
14.6
6.06
3.18
Etnylbenzene
0.27
3.77
3.47
3.89
Styrene
2.14
4.86
3.93
.14.1
*The quantity of each combustion product is reported in milligrams per gram of sample.
-------
POLYESTER
Several polyester film samples were sent to us by a county govern-
mental agency which was responsible for disposal of large quantities
of this material received at a landfill operation from a manufac-
turer in the area. Occasionally this film was received in such
large quantities that it could not be buried and was disposed of
by open burning. At this time landfill operators complained of
eye and skin irritation and complaints of a strong ouor in the air
were received from up to ten miles away.
Thermal analysis was not performed on these samples. When heated
at 5 C/min in the combustion furnace, the clear film melted down to
a pale yellow mass at about 200 C, and turned brown between 250
and 300 C. At 350 C it began bubbling and giving off a white fume
which condensed in the combustion tube and bag. The plastic was
completely combusted by 500 C.
By far the largest portion of the slowly heated polyester film
formed the white fume which condensed as a fine white powder. An
emission spectrum of the powder revealed no metallic elements,
indicating the powder to be organic rather than an inorganic in
nature. A mass spectrum of the powder taken using the solid probe
indicated a molecular weight of about 450 with a structure similar
to the following: Q ,—. Q 0 ,—. 0
H0-CH2-CH2-0"(i-<'o)-C-0-CH2-CH2-0-C-/O>-C-0-CH2-CH2-0H
This is probably a depolymerization product. Madorsky nas done a
more through analysis of similar residues.5
If some of the alcohol groups were replaced by aldehyde groups,
the above compound could produce irritation. However, no white
fume was noticed during the open burning so our attention was
turned to gaseous products. Infrared analysis of the combustion
gas showed bands due to carbon dioxide, carbon monoxide, methanol,
and acetaldehyde. Methane, ethane, and benzene were also identi-
fied by mass spectrometry after separation on either a Porapak Q
or Carbowax 20 M column. These were the only gaseous products
identified. Acetaldehyde is the only compound identified which
could cause the irritation and odor produced. Using the 5-C/min
heating rate and 100-cc/min air flow, approximately 0.1 milligrams
of acetaldehyde were found per gram of sample. However, this was
probably much more efficient combustion than open burning of a
bulky sample and it is possible that larger quantities of acetalde-
hyde could be produced.
This problem is typical of the disposal problems encountered with
bulky products in both incineration and open burning.
123
-------
SYNTHETIC FABRICS (DACRONR, ORLONR, NYLON)
The Flammable Fabrics Act includes a statement expressing the need
to know the nature and toxicity of the combustion products of fab-
rics, and yet the importance of this information seems to be dele-
gated to a secondary role. This is probably as it should be in
considering the flammability of clothing.. However, now that the
act has been extended to household and interior furnishings, more
attention should be given to this matter of the products of com-
bustion because our environment is rapidly changing from an "open"
to a "closed" nature. A few years ago we made provisions for
opening the windows of our homes and other buildings to enjoy the
fresh air. Now we go to increasingly greater efforts to seal
buildings to keep out the polluted air or air of the wrong
temperature. Likewise, in our modes of travel, for example, air
conditioned motor vehicles, aircraft flying at altitudes requiring
pressurization, and spacecraft traveling outside the atmosphere,
we require a tightly closed environment. In these situations, a
minor fire may not cause heat or burn damage to the occupant, but
can easily cause oxygen depletion, or the generation of toxic and
poisonous gases. We have already had examples in air travel where
passengers have survived an air crash and the resulting fire, only
to be killed by carbon monoxide and other products generated by
the fire.
Analysis of combustion products of fabrics was not within the scope
of the present grant and only preliminary thermal analyses have
been carried out,
Results
Thermal analysis. A differential thermal analysis record of Dac-
ron^ (Figure 54T shows a small endothermic peak at 260 C followed
by a large exothermic peak between 430 and 530 C. Thermogravimetric
analysis of the same sample (Figure 55) shows a two-step degra-
dation, the first step beginning about 330 C and ending at 410 C,
corresponding to an 85-percent weight loss. The remaining 15
percent of the sample is lost in an irregular fashion until the
plastic is completely combusted at about 500 C. Little interpre-
tation can be made of these records without some idea of products
generated, although it is likely that the DTA endotherm corresponds
to melting of the polymer and the DTA exotherm corresponds to the
major weight-loss of the sample. Combustion under the TGA condi-
tions seems to occur at slightly lower temperatures than at the
DTA conditions.
A similar situation exists for Orlon^ (Figure 56 and 57). There
is again a small DTA endotherm at 260 C and a large exothermic
peak between 390 and 525 C. The TGA record is parallel to that of
124
-------
T (°C)
Figure ^>k. Differential thermal analysis record of Dacron^
heated at 5 C/min in air
125
-------
T(°C)
Figure 55* Ir.ermogravimetric analysis record of Dacron^ heated at 10 C/min in air
-------
1'
En
100 200 300 400 500 600
T (°C)
Figure 56. Differential thermal analysis record of Orlon^
heated at 5 C/min in air
-------
T(°C)
Figure 57- T'nermogravimetric analysis record of Orlon^
heated at 10 c/ir.in in air
-------
QacronRbutis displaced about 10 C higher in temperature.
Nylon is similar to the other fabrics except that its DTA record
has an additional small exothermic peak at about 350 C and its TGA
record is displaced 10 C higher than OrlonR in temperature (Fig-
ures 58 and 59).
129
-------
100 200 300 400 500 600
T (°C)
Figure 58* Differencial thermal analysis record of nylon
heated at 5 c/min in air
-------
Figure 59T ;Thermogravimetric analysis record of nylon hee-tert at 10 C/min in air
-------
NATURAL PRODUCTS (WOOD AMD WOOL)
Introduction
Comparison of combustion products of plastics with combustion pro-
ducts of some materials replaced by plastics and with commonly-
incinerated materials is desirable in order to maintain the proper
perspective when interpreting combustion data. We chose wood and
wool with which to compare the carbon-hydrogen-oxygen-and the nitro-
gen-containing polymers, respectively. Our first priority, however,
was to study combustion products of plastics and work on these
natural products to date is of a very preliminary nature.
Results
Thermal analysis. Thermal analysis was not performed on these
samples.
Qualitative analysis. GC-MS analysis of the gaseous products of
combustion of wood on a five-foct-1ong, 1/4-in,-diameter Porapak Q
column (Figure 60) showed thirteen peaks in addition to air. Fif-
teen mass spectra were taken (sometimes more than ore spectrum per
peak, as numbered) and identifications are listed in Table 33.
Other compounds have not yet been identified. Gaseous products of
combustion of wool were analyzed by infrared spectroscopy, but
only carbon dioxide and carbon monoxide were positively identified.
Quantitative analysis. Analysis of wood (0.14 percent nitrogen)
combustion products for cyanide ion at three combustion conditions
(100-cc/mi n air, 5-and 50-C/min heating rates: and 100-cc/min air
plus 20-cc/min oxygen, 5-C/min heating rate) resulted in 0.3-0.4
milligrams cyanide per gram of sample.
132
-------
H
2_
-•—TIME
Figure 60. Chromatogram of wood combustion products on a Porapak Q column
-------
TABLE 33
IDENTIFICATION OF WOOD CHROMATOGRAM PEAKS
Scan
Number
Molecular
Wei ght
Identification
1
44
Carbon dioxide
2
28
Ethylene
3
30
Ethane
4
18
Water
5
--
Unidenti fied
6
42
Propylene
7
32
Methanol
44
Propane
8
32
Methanol
50
1,3-Butadiyne?
52
l-Buten-3-yne?
9
44
Acetaldehyde
10
60
n-Propanol
11
54
Butadiene
56"
Butene or Isobutene
12,13
56
Butene (Isomer?)
14
56
Butene (Isomer?)
15
58
Butane
68
Pentadiene
134
-------
ACKNOWLEDGMENTS
Any research of the scope and type reported here 1s generally de-
pendent on and Indebted to many peoplje other than the primary
researchers. This project is no exception, and the following
acknowledgments are in order:
First, our thanks to those students who assisted in the many rou-
tine tasks- especially to Tom Webster, Jill Robson, and Jim Kolton,
who helped during the last three years.
Second, we are indebted to those manufacturers who furnished us
samples, generally displaying an interest in the work, and some-
times indicating some trepidation in turning over data concerning
their samples, but still doing so. We appreciate the interest
shown us in our work by M. O'Mara, L, Crider, and W. E. McCormick
of the Goodrich Tire and Rubber Company, and thank them for the
many plastic samples furnished us. Likewise we are grateful to
all sample suppliers, including the Uniloy Division of the Hoover
Ball & Bearing Company, General Electric Company, E. I. DuPont de
Nemours and Company, 01 in Research Center, Monsanto Company,
Union Carbide & Chemical, Scott Paper Company, Mobil Chemical
Company, and Firestone Plastics Company.
Lastly, we are indebted to Louis Lefke, Dan Keller, Alvin Keene,
and E. Timothy Oppelt, formerly with the Solid Wastes Research
Laboratory of the U. S. Environmental Protection Agency. Especially
helpful 1n the latter phases of the program was Nancy Ulmer,
Project Officer, Solid Wastes Research Laboratory.
135
-------
REFERENCES
1. DeBell & Richardson, Inc. (A, J. Warner, C. H. Parker, and
B. Baum). Solid waste management of plastics, Washington,
Manufacturing Chemists Association, p. A-2, 1970.
2. Midwest Research Institute (A. Darnay and E. Franklin). Role
of packaging in solid wastes, 1966-1976. U.S. Department of
Health, Education, and Welfare publication No, 1855.
Washington, U.S. Government Printing Office.
3. Acrylo-based plastics debut as soft dririk bottles. Modern
Plastics, 47(5):62-63, May 1970.
4. The statistics for 1971, Modern Plastics, 49(1):41, Jan. 1972.
5. Madorsky, S, L. Thermal degradation of organic polymers.
New York, Interscience Publishers, 1964. 315 p.
6. Kaiser, E. R., and A. A. Carotti. Municipal incineration of
refuse with 2% and 4% additions of four plastics: polyethylene,
polystyrene, polyurethane, polyvinyl chloride. A report to the
Society of the Plastics Industry. New York, New York University,
1971.
7. Heimburg, R. W., e_t. al. Incineration of plastics found in
municipal wastes, Final report of Research Grant No. EC-00304
for the U.S. Environmental Protection Agency, Solid Waste
Research Laboratory. Syracuse, Syracuse University Corp.,
1972.
8. Miller, P. D., et. aj_. Fireside metal wastage in municipal
incinerators. Summary progress report of Research Grant No.
EC-00325 for the U.S. Environmental Protection Agency, Solid
Waste Research Laboratory. Columbus, Batelle Laboratories,
1971.
9. Boettner, E. A., and B. Weiss. An analytical system for iden-
tifying the volatile pyrolysis products of plastics. American
Industrial Hygiene Association Journal, 28:535-540, Nov.-Dec.
1967.
10. welcher, F. J., ed[. Standard methods of chemical analysis.
6th ed, Princeton, D, Van Nostrand Co., 1966. p. 1630,
11. Saltzman, B. E. Colorimetric microdetermination of nitrogen
dioxide in the atmosphere. Analytical Chemistry 26:1949-1955
(1954). " - - -
136
-------
12. Scoggins, M. W. Ultraviolet spectrophotometric determination
of sulfur dioxide. Analytical Chemistry, 42(9):1091 -1092,
Aug. 1970.
13. State of Michigan. House of Representatives. Banning the
sale or use of polyvinyl chloride bottles, House Bill no.
5486. June 25, 1971.
14. Boettner, E. A., G. Ball, and B. Weiss. Analysis of the
volatile combustion products of vinyl plastics. Journal of
Applied Polymer Science, 13:377-391, Feb. 1969.
15. Luther, H. and H. Kruger. The effect on the thermal decompo-
sition of polyvinyl chloride of the hydrogen chloride formed.
Translation from Kunststoffe, 56(2):74-79, Feb. 1966.
16. Braun, D. and M. Thallmaier. Measurement of the evolution of
hydrogen chloride from polyvinyl chloride at elevated
temperatures. Translation from Kunststoffe, 56(2):80-83,
Feb. 1966.
17. Vymazal, Z., B. Dolezel, and J. Stepek, Studies on the ther-
mal degradation of PVC and vinyl chloride copolymers. Trans-
lation from Kunststoffe, 56(2):86-91, Feb. 1966.
18. Clark, C. The burning issue of PVC disposal. Society of
Plastics Engineers Journal, 28:30-34, July 1972.
19. Cornish, H. and E. Abar. Toxicity of pyrolysis products of
vinyl plastics. Archives of Environmental Health, 19:15-21,
July 1969.
20. Boettner, E. A., principal investigator. Pyrolytic products
of plastics: analysis and toxicity. Final report of Research
Grant no. 0H-00148 (also UI-00485 and EC-00232) for the U.S.
Public Health Service. Ann Arbor, University of Michigan,
July 1970.
21. Gross, S., ed. Modern plastics encyclopedia, v. 48. New York,
McGraw-Hill, 1971. p. 51, 105-106.
22. Davis, A. Thermal stability of Dolysulfone, Makromolecular
Chemie, 128:242, 1967.
23. Hale, W. Poly (aryl) ethers by nucleophilic aromatic sub-
stitution. Part II: Thermal stability. Journal of Polymer
Science, Part A-l, 5:2399-2414, 1967.
24. Zapp, J. A. Hazards of isocyanates in polyurethane foam
plastic production. A.M.A. Archives of Industrial Health,
137
-------
15:324, 1957,
25. Seader, J., et. a_L Analysis of volatile combustion products
and a study o? their toxicological effects. Polymer
Engineering and Science, 12(2):125-133, Mar. 1972.
26. Takeuchi, T., S. Tsuge, and T. Okumoto. Identification and
analysis of urethane foams by pyrolysis-gas chromatography.
Journal of Gas Chromatography, 6:542-547, Nov. 1968.
27. Bishop, D. and D. Smith. Combined pyrolysis and radiochemical
gas chromatography for studying the thermal degradation of
epoxide resins and polyimides. Part II: Degradation of
polyimides. Journal of Applied Polymer Science, 14:345-354,
1970.
28. Heacock, J. F. and C. E. Berr. Polyimides-new high-tempera-
ture polymers: H-film, a polypyromellitimide film. SPE
Transactions, 2:105-110, Apr. 1965.
29. Monsanto issues safe burning data on LopacR plastic containers.
Solid Waste Report, 2:60, Apr, 1971.
30. Idol, J. D. BarexR 210, a resin for plastic bottles which
can be safely and easily incinerated. Abstract of paper
presented at the Regional Technical Conference of the Society
of Plastics Engineers, Inc., Cherry Hill Inn, Cherry Hill,
New Jersey, Oct. 27-28, 1970.
31. Burns, R. and E. W. Orrell. A thermal analytical study of
phenol-formaldehyde resins. Journal of Materials Science,
2(1) :72-77, Jan. 1967.
32. Shebozaki, Y. et^ al. Thermogravimetric analysis of phenolic
resins. Part I: Thermal degradation of phenol-formaldehyde
and ct- and 3-naphthol-formaldehyde resins. Chemical Abstracts,
71:32 (abstract 71278), 1969.
33. TRW Systems Group, (M. E. Banks, W. D. Lusk, and R. S.r
Ottinger) New chemical concepts for utilization of waste
plastics. Final report of Contract No. PH 86-68-206 for the
U.S. Environmental Protection Agency, 1971.
34. Davis, A. and J. H. Golden. Degradation of polycarbonates.
Part III. Viscometric study of thermally induced chain
scission. Makromolecular Chemie, 78:16, 1964.
35. Davis, A. and J. H. Golden. Competition between scission and
cross-linking processes in the thermal degradation of a poly-
carbonate. Nature, 206 (4982):397, 1965.
138
-------
36, Lee, L, H, Mechanisms of thermal degradation of phenolic
condensation polymers. Part I: Studies on the thermal
stability of polycarbonate, Journal of Applied Polymer
Science: Part A, 2:2859-2873, 1964,
37, Flashman, B. P., and P. Shenian. Polyphenylene oxide - a new
engineering material. Rubber Plastics Age, 47:383, Apr.
1966.
38. Gowan, A. C. More data on polyphenylene oxide. Hydrocarbon
Processing and Petroleum Refiner, 45:180, Nov. 1966.
39. Cox, J. M., B. A. Wright, and W. W. Wright. Thermal degrad-
ation of poly (phenylene oxides). Journal of Applied Polymer
Science, 9:513, Feb, 1965.
139
-------
PUBLICATIONS RESULTING FROM RESEARCH
1. Ball, G., B. Weiss, and E. A, Boettner. Analysis of the
volatile combustion products of polyphenylene oxide plastics.
American Industrial Hygiene Association Journal, 31:572-587.
Sept.-Oct. 1970.
2. Ball, G., and E. A. Boettner. Volatile combustion products of
polycarbonate and polysulfone. Journal of Applied Polymer
Science, 16:855-853. 1972.
3. Boettner, E. A. and G. Ball. Combustion products of plastics
and their contribution to incineration problems. AIChE
Symposium Series. 68:13-20. 1972.
140
-------
THE FOLLOWING PAGES ARE DUPLICATES OF
ILLUSTRATIONS APPEARING ELSEWHERE IN THIS
REPORT. THEY HAVE BEEN REPRODUCED HERE BY
A DIFFERENT METHOD TO PROVIDE BETTER DETAIL.
-------
-------
-------
------- |